Development of a dual-energy spectral CT based nomogram for the preoperative discrimination of mutated and wild-type KRAS in patients with colorectal cancer

Deep learning of pretreatment multiphase CT images for predicting response to lenvatinib and immune checkpoint inhibitors in unresectable hepatocellular carcinoma

Objectives

Combination therapy of lenvatinib and immune checkpoint inhibitors (CLICI) has emerged as a promising approach for managing unresectable hepatocellular carcinoma (HCC). However, the response to such treatment is observed in only a subset of patients, underscoring the pressing need for reliable methods to identify potential responders.

Materials & methods

This was a retrospective analysis involving 120 patients with unresectable HCC. They were divided into training (n = 72) and validation (n = 48) cohorts. We developed an interpretable deep learning model using multiphase computed tomography (CT) images to predict whether patients will respond or not to CLICI treatment, based on the Response Evaluation Criteria in Solid Tumors, version 1.1 (RECIST v1.1). We evaluated the models' performance and analyzed the impact of each CT phase. Critical regions influencing predictions were identified and visualized through heatmaps.

Results

The multiphase model outperformed the best biphase and uniphase models, achieving an area under the curve (AUC) of 0.802 (95% CI = 0.780-0.824). The portal phase images were found to significantly enhance the model's predictive accuracy. Heatmaps identified six critical features influencing treatment response, offering valuable insights to clinicians. Additionally, we have made this model accessible via a web server at http://uhccnet.com/ for ease of use.

Conclusions

The integration of multiphase CT images with deep learning-generated heatmaps for predicting treatment response provides a robust and practical tool for guiding CLICI therapy in patients with unresectable HCC.

Construction of prediction model for KRAS mutation status of colorectal cancer based on CT radiomics

BACKGROUND

In this study, we used computed tomography (CT)-based radiomics signatures to predict the mutation status of KRAS in patients with colorectal cancer (CRC) and to identify the phase of radiomics signature with the most robust and high performance from triphasic enhanced CT.

METHODS

This study involved 447 patients who underwent KRAS mutation testing and preoperative triphasic enhanced CT. They were categorized into training (n = 313) and validation cohorts (n = 134) in a 7:3 ratio. Radiomics features were extracted using triphasic enhanced CT imaging. The Boruta algorithm was used to retain the features closely associated with KRAS mutations. The Random Forest (RF) algorithm was used to develop radiomics, clinical, and combined clinical-radiomics models for KRAS mutations. The receiver operating characteristic curve, calibration curve, and decision curve were used to evaluate the predictive performance and clinical usefulness of each model.

RESULTS

Age, CEA level, and clinical T stage were independent predictors of KRAS mutation status. After rigorous feature screening, four arterial phase (AP), three venous phase (VP), and seven delayed phase (DP) radiomics features were retained as the final signatures for predicting KRAS mutations. The DP models showed superior predictive performance compared to AP or VP models. The clinical-radiomics fusion model showed excellent performance, with an AUC, sensitivity, and specificity of 0.772, 0.792, and 0.646 in the training cohort, and 0.755, 0.724, and 0.684 in the validation cohort, respectively. The decision curve showed that the clinical-radiomics fusion model had more clinical practicality than the single clinical or radiomics model in predicting KRAS mutation status.

CONCLUSION

The clinical-radiomics fusion model, which combines the clinical and DP radiomics model, has the best predictive performance for predicting the mutation status of KRAS in CRC, and the constructed model has been effectively verified by an internal validation cohort.

Predictive modeling based on tumor spectral CT parameters and clinical features for postoperative complications in patients undergoing colon resection for cancer

BACKGROUND

Colon cancer is a particularly prevalent malignancy that produces postoperative complications (POCs). However, limited imaging modality exists on the accurate diagnosis of POCs. The purpose of this study was therefore to construct a model combining tumor spectral CT parameters and clinical features to predict POCs before surgery in colon cancer.

METHODS

This retrospective study included 85 patients who had preoperative abdominal spectral CT scans and underwent radical colon cancer resection at our institution. The patients were divided into two groups based on the absence (no complication/grade I) or presence (grades II-V) of POCs according to the Clavien-Dindo grading system. The visceral fat areas (VFA) of patients were semi-automatically outlined and calculated on L3-level CT images using ImageJ software. Clinical features and tumor spectral CT parameters were statistically compared between the two groups. A combined model of spectral CT parameters and clinical features was established by stepwise regression to predict POCs in colon cancer. The diagnostic performance of the model was evaluated using the receiver operating characteristic (ROC) curve, including area under the curve (AUC), sensitivity, and specificity.

RESULTS

Twenty-seven patients with POCs and 58 patients without POCs were included in this study. MonoE40keV-VP and VFA were independent predictors of POCs. The combined model based on predictors yielded an AUC of 0.84 (95% CI: 0.74-0.91), with a sensitivity of 77.8% and specificity of 87.9%.

CONCLUSIONS

The model combining MonoE40keV-VP and VFA can predict POCs before surgery in colon cancer and provide a basis for individualized management plans.

CRITICAL RELEVANCE STATEMENT

The model combining MonoE40keV-VP and visceral fat area can predict postoperative complications before surgery in colon cancer and provide a basis for individualized management plans.

KEY POINTS

• Visceral fat area and MonoE40keV-VP were independent predictors of postoperative complications in colon cancer. • The combined model yielded a high AUC, sensitivity, and specificity in predicting postoperative complications. • The combined model was superior to the single visceral fat area or MonoE40keV-VP in predicting postoperative complications.

Overview of serum and tissue markers in colorectal cancer: a primer for radiologists

Serum and tissue tumor markers provide crucial information in the diagnosis, treatment, and follow-up of colorectal cancers. Tissue tumor markers are increasingly used for determination of targeted chemotherapy planning based on genotyping of tumor cells. Recently, plasma-based technique of liquid biopsy is being evaluated for providing tumor biomarkers in the management of colorectal cancer. Tumor markers are commonly used in conjunction with imaging during initial staging, treatment determination, response assessment, and determination of recurrence or metastatic disease. Knowledge of tumor markers and their association with radiological findings is thus crucial for radiologists. Additionally, various novel imaging techniques are being evaluated as potential noninvasive imaging biomarkers to predict tumor genotypes, features, and tumor response. We review and discuss the potential role of these newer imaging techniques.

Effect of spectral CT on tumor microvascular angiogenesis in renal cell carcinoma

Background

To examine the value of energetic-spectrum computed tomography (spectral CT) quantitative parameters in renal cell carcinoma (RCC) microvascular angiogenesis.

Methods

The authors evaluated 32 patients with pathologically confirmed RCC who underwent triple-phase contrast-enhanced CT with spectral CT imaging mode from January 2017 to December 2019. Quantitative parameters include parameters derived from iodine concentration (IC) and water concentration (WC) of 120 keV monochromatic images. All specimens were evaluated including the microvascular density (MVD), microvascular area (MVA) and so on. The correlation between IC and WC (including average values and random values) with microvascular parameters were analyzed with Pearson or Spearman rank correlation coefficients.

Results

The MVD of all tumors was 26.00 (15.00–43.75) vessels per field at × 400 magnification. The MVD of RCC correlated positively with the mean IC, mean WC, mean NWC, mean NIC, random IC, random NIC in renal cortical phase, WCD₁, WCD₂, NWCD₂ and ICD₁ (Spearman rank correlation coefficients, r range, 0.362–0.533; all p < 0.05). The MVA of all tumors was (16.16 ± 8.98) % per field at × 400 magnification. The MVA of RCC correlated positively with the mean IC, mean WC, mean NWC, mean NIC, random IC, random NIC in renal cortical, mean WC and mean NWC in renal parenchymal phase, WCD₁, WCD₂, WCD₃, NWCD₂, and NWCD₃ (Pearson or Spearman rank correlation coefficients, r range, 0.357–0.576; all p < 0.05). Microvascular grading correlated positively with the mean NWC, mean NIC and random NIC in renal cortical phase, mean NWC in renal parenchymal phase, NWCD₂, WCD₃, NWCD₃, NICD₂ and NICD₃ (Spearman rank correlation coefficients, r range, 0.367–0.520; all p < 0.05). As for tumor diameter (55.19 ± 19.15), μm, only NWCD₃ was associated with it (Spearman rank correlation coefficients, r = 0.388; p < 0.05).

Conclusions

ICD and WCD of spectral CT have a potential for evaluating RCC microvascular angiogenesis. MVD, MVA and microvascular grade showed moderate positive correlation with ICD and WCD. ICD displayed more relevant than that of WCD. The parameters of renal cortical phase were the best in three phases. NICD and NWCD manifested stronger correlation with microvascular parameters than that of ICD and WCD.

Advances in radiological staging of colorectal cancer

The role of imaging in clinically staging colorectal cancer has grown substantially in the 21^st century with more widespread availability of multi-row detector computed tomography (CT), high-resolution magnetic resonance imaging (MRI) with diffusion weighted imaging (DWI), and integrated positron-emission tomography (PET)/CT. In contrast to staging many other cancers, increasing colorectal cancer stage does not highly correlate with survival. As has been the case previously, clinical practice incorporates advances in staging and it is used to guide therapy before adoption into international staging guidelines. Emerging imaging techniques show promise to become part of future staging standards.

Predicting Microsatellite Instability Status in Colorectal Cancer Based on Triphasic Enhanced Computed Tomography Radiomics Signatures: A Multicenter Study

Background

This study aimed to develop and validate a computed tomography (CT)-based radiomics model to predict microsatellite instability (MSI) status in colorectal cancer patients and to identify the radiomics signature with the most robust and high performance from one of the three phases of triphasic enhanced CT.

Methods

In total, 502 colorectal cancer patients with preoperative contrast-enhanced CT images and available MSI status (441 in the training cohort and 61 in the external validation cohort) were enrolled from two centers in our retrospective study. Radiomics features of the entire primary tumor were extracted from arterial-, delayed-, and venous-phase CT images. The least absolute shrinkage and selection operator method was used to retain the features closely associated with MSI status. Radiomics, clinical, and combined Clinical Radiomics models were built to predict MSI status. Model performance was evaluated by receiver operating characteristic curve analysis.

Results

Thirty-two radiomics features showed significant correlation with MSI status. Delayed-phase models showed superior predictive performance compared to arterial- or venous-phase models. Additionally, age, location, and carcinoembryonic antigen were considered useful predictors of MSI status. The Clinical Radiomics nomogram that incorporated both clinical risk factors and radiomics parameters showed excellent performance, with an AUC, accuracy, and sensitivity of 0.898, 0.837, and 0.821 in the training cohort and 0.964, 0.918, and 1.000 in the validation cohort, respectively.

Conclusions

The proposed CT-based radiomics signature has excellent performance in predicting MSI status and could potentially guide individualized therapy.

Highly heterogeneous-related genes of triple-negative breast cancer: potential diagnostic and prognostic biomarkers

Background

Triple-negative breast cancer (TNBC) is a highly heterogeneous subtype of breast cancer, showing aggressive clinical behaviors and poor outcomes. It urgently needs new therapeutic strategies to improve the prognosis of TNBC. Bioinformatics analyses have been widely used to identify potential biomarkers for facilitating TNBC diagnosis and management.

Methods

We identified potential biomarkers and analyzed their diagnostic and prognostic values using bioinformatics approaches. Including differential expression gene (DEG) analysis, Receiver Operating Characteristic (ROC) curve analysis, functional enrichment analysis, Protein-Protein Interaction (PPI) network construction, survival analysis, multivariate Cox regression analysis, and Non-negative Matrix Factorization (NMF).

Results

A total of 105 DEGs were identified between TNBC and other breast cancer subtypes, which were regarded as heterogeneous-related genes. Subsequently, the KEGG enrichment analysis showed that these genes were significantly enriched in ‘cell cycle’ and ‘oocyte meiosis’ related pathways. Four (FAM83B, KITLG, CFD and RBM24) of 105 genes were identified as prognostic signatures in the disease-free interval (DFI) of TNBC patients, as for progression-free interval (PFI), five genes (FAM83B, EXO1, S100B, TYMS and CFD) were obtained. Time-dependent ROC analysis indicated that the multivariate Cox regression models, which were constructed based on these genes, had great predictive performances. Finally, the survival analysis of TNBC subtypes (mesenchymal stem-like [MSL] and mesenchymal [MES]) suggested that FAM83B significantly affected the prognosis of patients.

Conclusions

The multivariate Cox regression models constructed from four heterogeneous-related genes (FAM83B, KITLG, RBM24 and S100B) showed great prediction performance for TNBC patients’ prognostic. Moreover, FAM83B was an important prognostic feature in several TNBC subtypes (MSL and MES). Our findings provided new biomarkers to facilitate the targeted therapies of TNBC and TNBC subtypes.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12885-021-08318-1.

Development and Validation of a Contrast-Enhanced CT-Based Radiomics Nomogram for Prediction of Therapeutic Efficacy of Anti-PD-1 Antibodies in Advanced HCC Patients

Background

There is no study accessible now assessing the prognostic aspect of radiomics for anti-PD-1 therapy for patients with HCC.

Aim

The aim of this study was to develop and validate a radiomics nomogram by incorporating the pretreatment contrast-enhanced Computed tomography (CT) images and clinical risk factors to estimate the anti-PD-1 treatment efficacy in Hepatocellular Carcinoma (HCC) patients.

Methods

A total of 58 patients with advanced HCC who were refractory to the standard first-line of therapy, and received PD-1 inhibitor treatment with Toripalimab, Camrelizumab, or Sintilimab from 1st January 2019 to 31 July 2020 were enrolled and divided into two sets randomly: training set (n = 40) and validation set (n = 18). Radiomics features were extracted from non-enhanced and contrast-enhanced CT scans and selected by using the least absolute shrinkage and selection operator (LASSO) method. Finally, a radiomics nomogram was developed based on by univariate and multivariate logistic regression analysis. The performance of the nomogram was evaluated by discrimination, calibration, and clinical utility.

Results

Eight radiomics features from the whole tumor and peritumoral regions were selected and comprised of the Fusion Radiomics score. Together with two clinical factors (tumor embolus and ALBI grade), a radiomics nomogram was developed with an area under the curve (AUC) of 0.894 (95% CI, 0.797–0.991) and 0.883 (95% CI, 0.716–0.998) in the training and validation cohort, respectively. The calibration curve and decision curve analysis (DCA) confirmed that nomogram had good consistency and clinical usefulness.

Conclusions

This study has developed and validated a radiomics nomogram by incorporating the pretreatment CECT images and clinical factors to predict the anti-PD-1 treatment efficacy in patients with advanced HCC.