Multiparametric MRI model to predict molecular subtypes of breast cancer using Shapley additive explanations interpretability analysis.

PURPOSE: The purpose of this study was to assess the predictive performance of multiparametric magnetic resonance imaging (MRI) for molecular subtypes and interpret features using SHapley Additive exPlanations (SHAP) analysis. MATERIAL AND METHODS: Patients with breast cancer who underwent pre-treatment MRI (including ultrafast dynamic contrast-enhanced MRI, magnetic resonance spectroscopy, diffusion kurtosis imaging and intravoxel incoherent motion) were recruited between February 2019 and January 2022. Thirteen semantic and thirteen multiparametric features were collected and the key features were selected to develop machine-learning models for predicting molecular subtypes of breast cancers (luminal A, luminal B, triple-negative and HER2-enriched) by using stepwise logistic regression. Semantic model and multiparametric model were built and compared based on five machine-learning classifiers. Model decision-making was interpreted using SHAP analysis. RESULTS: A total of 188 women (mean age, 53 ± 11 [standard deviation] years; age range: 25-75 years) were enrolled and further divided into training cohort (131 women) and validation cohort (57 women). XGBoost demonstrated good predictive performance among five machine-learning classifiers. Within the validation cohort, the areas under the receiver operating characteristic curves (AUCs) for the semantic models ranged from 0.693 (95% confidence interval [CI]: 0.478-0.839) for HER2-enriched subtype to 0.764 (95% CI: 0.681-0.908) for luminal A subtype, inferior to multiparametric models that yielded AUCs ranging from 0.771 (95% CI: 0.630-0.888) for HER2-enriched subtype to 0.857 (95% CI: 0.717-0.957) for triple-negative subtype. The AUCs between the semantic and the multiparametric models did not show significant differences (P range: 0.217-0.640). SHAP analysis revealed that lower iAUC, higher kurtosis, lower D*, and lower kurtosis were distinctive features for luminal A, luminal B, triple-negative breast cancer, and HER2-enriched subtypes, respectively. CONCLUSION: Multiparametric MRI is superior to semantic models to effectively predict the molecular subtypes of breast cancer.

Early Identification of Pathologic Complete Response to Neoadjuvant Chemotherapy Using Multiphase DCE-MRI by Siamese Network in Breast Cancer: A Longitudinal Multicenter Study.

BACKGROUND: Siamese network (SN) using longitudinal DCE-MRI for pathologic complete response (pCR) identification lack a unified approach to phases selection. PURPOSE: To identify pCR in early-stage NAC, using SN with longitudinal DCE-MRI and introducing IPS for phases selection. STUDY TYPE: Multicenter, longitudinal. POPULATION: Center A: 162 female patients (50.63 ± 8.41 years) divided 7:3 into training and internal validation cohorts. Center B: 61 female patients (50.08 ± 7.82 years) were used as an external validation cohort. FIELD STRENGTH/SEQUENCE: Center A: single vendor 3.0 T with a compressed-sensing volume interpolated breath-hold examination sequence. Center B: single vendor 1.5 T with volume interpolated breath-hold examination sequence. ASSESSMENT: Patients underwent DCE-MRI before and after two NAC cycles, with tumor regions of interest (ROI) manually delineated. Histopathology was the reference for pCR identification. Models developed included a clinical one, four SN models based on IPS-selected phases, and integrated models combining clinical and SN features. STATISTICAL TESTS: Model performance was evaluated using the area under the receiver operating characteristic curve (AUC). The DeLong test was used to compare AUCs. Net reclassification improvement and integrated discrimination improvement (IDI) tests were employed for performance comparison. P < 0.05 was considered significant. RESULTS: In internal and external validation cohorts, the clinical model showed AUCs of 0.760 and 0.718. SN and integrated models, with increasing phases via IPS, achieved AUCs ranging from 0.813 to 0.951 and 0.818 to 0.922. Notably, SN-3 and integrated-3 and integrated-4 outperformed the clinical model. However, input phases beyond 20% did not significantly enhance performance (IDI test: SN-4 vs. SN-3, P = 0.314 and 0.630; integrated-4 vs. integrated-3, P = 0.785 and 0.709). DATA CONCLUSION: The longitudinal multiphase DCE-MRI based on the SN demonstrates promise for identifying pCR in breast cancer. EVIDENCE LEVEL: 1 TECHNICAL EFFICACY: Stage 4.

Discriminant ability of the shock index, modified shock index, and reverse shock index multiplied by the Glasgow coma scale on mortality in adult trauma patients: a PATOS retrospective cohort study.

BACKGROUND: The shock index (SI) predicts short-term mortality in trauma patients. Other shock indices have been developed to improve discriminant accuracy. The authors examined the discriminant ability of the SI, modified SI (MSI), and reverse SI multiplied by the Glasgow Coma Scale (rSIG) on short-term mortality and functional outcomes. METHODS: The authors evaluated a cohort of adult trauma patients transported to emergency departments. The first vital signs were used to calculate the SI, MSI, and rSIG. The areas under the receiver operating characteristic curves and test results were used to compare the discriminant performance of the indices on short-term mortality and poor functional outcomes. A subgroup analysis of geriatric patients with traumatic brain injury, penetrating injury, and nonpenetrating injury was performed. RESULTS: A total of 105 641 patients (49±20 years, 62% male) met the inclusion criteria. The rSIG had the highest areas under the receiver operating characteristic curve for short-term mortality (0.800, CI: 0.791-0.809) and poor functional outcome (0.596, CI: 0.590-0.602). The cutoff for rSIG was 18 for short-term mortality and poor functional outcomes with sensitivities of 0.668 and 0.371 and specificities of 0.805 and 0.813, respectively. The positive predictive values were 9.57% and 22.31%, and the negative predictive values were 98.74% and 89.97%. rSIG also had better discriminant ability in geriatrics, traumatic brain injury, and nonpenetrating injury. CONCLUSION: The rSIG with a cutoff of 18 was accurate for short-term mortality in Asian adult trauma patients. Moreover, rSIG discriminates poor functional outcomes better than the commonly used SI and MSI.

Development and validation of a radiogenomics model to predict axillary lymph node metastasis in breast cancer integrating MRI with transcriptome data: A multicohort study.

Introduction: To develop and validate a radiogenomics model for predicting axillary lymph node metastasis (ALNM) in breast cancer compared to a genomics and radiomics model. Methods: This retrospective study integrated transcriptomic data from The Cancer Genome Atlas with matched MRI data from The Cancer Imaging Archive for the same set of 111 patients with breast cancer, which were used as the training and testing groups. Fifteen patients from one hospital were enrolled as the external validation group. Radiomics features were extracted from dynamic contrast-enhanced (DCE)-MRI of breast cancer, and genomics features were derived from differentially expressed gene analysis of transcriptome data. Boruta was used for genomics and radiomics data dimension reduction and feature selection. Logistic regression was applied to develop genomics, radiomics, and radiogenomics models to predict ALNM. The performance of the three models was assessed by receiver operating characteristic curves and compared by the Delong test. Results: The genomics model was established by nine genomics features, and the radiomics model was established by three radiomics features. The two models showed good discrimination performance in predicting ALNM in breast cancer, with areas under the curves (AUCs) of 0.80, 0.67, and 0.52 for the genomics model and 0.72, 0.68, and 0.71 for the radiomics model in the training, testing and external validation groups, respectively. The radiogenomics model integrated with five genomics features and three radiomics features had a better performance, with AUCs of 0.84, 0.75, and 0.82 in the three groups, respectively, which was higher than the AUC of the radiomics model in the training group and the genomics model in the external validation group (both P < 0.05). Conclusion: The radiogenomics model combining radiomics features and genomics features improved the performance to predict ALNM in breast cancer.

Association between the time to definitive care and trauma patient outcomes: every minute in the golden hour matters.

PURPOSE: This study examined the association between lapsed time and trauma patients, suggesting that a shorter time to definitive care leads to a better outcome. METHODS: We used the Pan-Asian Trauma Outcome Study registry to analyze a retrospective cohort of 963 trauma patients who received surgical intervention or transarterial embolization within 2 h of injury in Asian countries between January 2016 and December 2020. Exposure measurement was recorded every 30 min from injury to definitive care. The 30 day mortality rate and functional outcome were studied using the Modified Rankin Scale ratings of 0-3 vs 4-6 for favorable vs poor functional outcomes, respectively. Subgroup analyses of different injury severities and patterns were performed. RESULTS: The mean time from injury to definitive care was 1.28 ± 0.69 h, with cases categorized into the following subgroups: < 30, 30-60, 60-90, and 90-120 min. For all patients, a longer interval was positively associated with the 30 day mortality rate (p = 0.053) and poor functional outcome (p < 0.05). Subgroup analyses showed the same association in the major trauma (n = 321, p < 0.05) and torso injury groups (n = 388, p < 0.01) with the 30 day mortality rate and in the major trauma (p < 0.01), traumatic brain injury (n = 741, p < 0.05), and torso injury (p < 0.05) groups with the poor functional outcome. CONCLUSION: Even within 2 h, a shorter time to definitive care is positively associated with patient survival and functional outcome, especially in the subgroups of major trauma and torso injury.

Tumor Stage-Based Gross Tumor Volume of Resectable Esophageal Squamous Cell Carcinoma Measured on CT: Association With Early Recurrence After Esophagectomy.

OBJECTIVE: To investigate relationship of tumor stage-based gross tumor volume (GTV) of esophageal squamous cell carcinoma (ESCC) measured on computed tomography (CT) with early recurrence (ER) after esophagectomy. MATERIALS AND METHODS: Two hundred and four consecutive patients with resectable ESCC including 159 patients enrolled in the training cohort (TC) and 45 patients in validation cohort (VC) underwent contrast-enhanced CT less than 2 weeks before esophagectomy. GTV was retrospectively measured by multiplying sums of all tumor areas by section thickness. For the TC, univariate and multivariate analyses were performed to determine factors associated with ER. Mann-Whitney U test was conducted to compare GTV in patients with and without ER. Receiver operating characteristic (ROC) analysis was performed to determine if tumor stage-based GTV could predict ER. For the VC, unweighted Cohen's Kappa tests were used to evaluate the performances of the previous ROC predictive models. RESULTS: ER occurred in 63 of 159 patients (39.6%) in the TC. According to the univariate analysis, histologic differentiation, cT stage, cN stage, and GTV were associated with ER after esophagectomy (all P-values < 0.05). Multivariate analysis revealed that cT stage and GTV were independent risk factors with hazard ratios of 3.382 [95% confidence interval (CI): 1.533-7.459] and 1.222 (95% CI: 1.125-1.327), respectively (all P-values < 0.05). Mann-Whitney U tests showed that GTV could help differentiate between ESCC with and without ER in stages cT1-4a, cT2, and cT3 (all P-values < 0.001), and the ROC analysis demonstrated the corresponding cutoffs of 13.31, 17.22, and 17.83 cm3 with areas under the curve of more than 0.8, respectively. In the VC, the Kappa tests validated that the ROC predictive models had good performances for differentiating between ESCC with and without ER in stages cT1-4a, cT2, and cT3 with Cohen k of 0.696 (95% CI, 0.498-0.894), 0.733 (95% CI, 0.386-1.080), and 0.862 (95% CI, 0.603-1.121), respectively. CONCLUSION: GTV and cT stage can be independent risk factors of ER in ESCC after esophagectomy, and tumor stage-based GTV measured on CT can help predict ER.

Contrast-enhanced CT radiomics features to predict recurrence of locally advanced oesophageal squamous cell cancer within 2 years after trimodal therapy: A case-control study.

ABSTRACT: Radiomics transforms the medical images into high-dimensional quantitative features and provides potential information about tumor phenotypes and heterogeneity. We conducted a retrospective analysis to explore and validate radiomics model based on contrast-enhanced computed tomography (CECT) to predict recurrence of locally advanced oesophageal squamous cell cancer (SCC) within 2 years after trimodal therapy. This study collected CECT and clinical data of consecutive 220 patients with pathology-confirmed locally advanced oesophageal SCC (154 in the training cohort and 66 in the validation cohort). Univariate statistical test and the least absolute shrinkage and selection operator method were performed to select the optimal radiomics features. Logistic regression was conducted to build radiomics model, clinical model, and combined model of both the radiomics and clinical features. Predictive performance was judged by the area under receiver operating characteristics curve (AUC), accuracy, and F1-score in the training and validation cohorts. Ten optimal radiomics features and/or 7 clinical features were selected to build radiomics model, clinical model, and the combined model. The integrated model of radiomics and clinical features was superior to radiomics model or clinical model in predicting recurrence of locally advanced oesophageal SCC within 2 years in the training (AUC: 0.879 vs 0.815 or 0.763; accuracy: 0.844 vs 0.773 or 0.740; and F1-score: 0.886 vs 0.839 or 0.815, respectively) and validation (AUC: 0.857 vs 0.720 or 0.750; accuracy: 0.788 vs 0.700 or 0.697; and F1-score: 0.851 vs 0.800 or 0.787, respectively) cohorts. The combined model of radiomics and clinical features shows better performance than the radiomics or clinical model to predict the recurrence of locally advanced oesophageal SCC within 2 years after trimodal therapy.

Application of contrast-enhanced CT radiomics in prediction of early recurrence of locally advanced oesophageal squamous cell carcinoma after trimodal therapy.

BACKGROUND: Early recurrence of oesophageal squamous cell carcinoma (SCC) is defined as recurrence after surgery within 1 year, and appears as local recurrence, distant recurrence, and lymph node positive and disseminated recurrence. Contrast-enhanced computed tomography (CECT) is recommended for diagnosis of primary tumor and initial staging of oesophageal SCC, but it cannot be used to predict early recurrence. It is reported that radiomics can help predict preoperative stages of oesophageal SCC, lymph node metastasis before operation, and 3-year overall survival of oesophageal SCC patients following chemoradiotherapy by extracting high-throughput quantitative features from CT images. This study aimed to develop models based on CT radiomics and clinical features of oesophageal SCC to predict early recurrence of locally advanced cancer. METHODS: We collected electronic medical records and image data of 197 patients with confirmed locally advanced oesophageal SCC. These patients were randomly allocated to 137 patients in the training cohort and 60 in the test cohort. 352 radiomics features were extracted by delineating region-of-interest (ROI) around the lesion on CECT images and clinical signature was generated by medical records. The radiomics model, clinical model, the combined model of radiomics and clinical features were developed by radiomics features and/or clinical characteristics. Predicting performance of the three models was assessed with area under receiver operating characteristic curve (AUC), accuracy and F-1 score. RESULTS: Eleven radiomics features and/or six clinical signatures were selected to build prediction models related to recurrence of locally advanced oesophageal SCC after trimodal therapy. The AUC of integration of radiomics and clinical models was better than that of radiomics or clinical model for the training cohort (0.821 versus 0.754 or 0.679, respectively) and for the validation cohort (0.809 versus 0.646 or 0.658, respectively). Integrated model of radiomics and clinical features showed good performance in predicting early recurrence of locally advanced oesophageal SCC for both the training and validation cohorts (accuracy = 0.730 and 0.733, and F-1score = 0.730 and 0.778, respectively). CONCLUSIONS: The integrated model of CECT radiomics and clinical features may be a potential imaging biomarker to predict early recurrence of locally advanced oesophageal SCC after trimodal therapy.

CT radiomics features to predict lymph node metastasis in advanced esophageal squamous cell carcinoma and to discriminate between regional and non-regional lymph node metastasis: a case control study.

BACKGROUND: Prediction of lymph node status in esophageal squamous cell carcinoma (ESCC) is critical for clinical decision making. In clinical practice, computed tomography (CT) has been frequently used to assist in the preoperative staging of ESCC. Texture analysis can provide more information to reflect potential biological heterogeneity based on CT. A nomogram for the preoperative diagnosis of lymph node metastasis in patients with resectable ESCC has been previously developed. However, to the best of our knowledge, no reports focus on developing CT radiomics features to discriminate ESCC patients with regional lymph node metastasis (RLNM) and non-regional lymph node metastasis (NRLNM). We, therefore, aimed to develop CT radiomics models to predict lymph node metastasis (LNM) in advanced ESCC and to discriminate ESCC between RLNM and NRLNM. METHODS: This study enrolled 334 patients with pathologically confirmed advanced ESCC, including 152 patients without LNM and 182 patients with LNM, and 103 patients with RLNM and 79 patients NRLNM. Radiomics features were extracted from CT data for each patient. The least absolute shrinkage and selection operator (LASSO) model and independent samples t-tests or Mann-Whitney U tests were exploited for dimension reduction and selection of radiomics features. Optimal radiomics features were chosen using multivariable logistic regression analysis. The discriminating performance was assessed by area under the receiver operating characteristic curve (AUC) and accuracy. RESULTS: The radiomics features were developed based on multivariable logistic regression and were significantly associated with LNM status in both the training and validation cohorts (P<0.001). The radiomics models could differentiate between patients with and without LNM (AUC =0.79 and 0.75, and accuracy =0.75 and 0.71 in the training and validation cohorts, respectively). In patients with LNM, the radiomics features could effectively differentiate between RLNM and NRLNM (AUC =0.98 and 0.95, and accuracy =0.94 and 0.83 in the training and validation cohorts, respectively). CONCLUSIONS: CT radiomics features could help predict the LNM status of advanced ESCC patients and effectively discriminate ESCC between RLNM and NRLNM.

Changes in CT manifestations and RT-PCR testings of the coronavirus disease 2019 until recovery in patients with afferent infection vs. second-generation infection outside the original city (Wuhan): An observational study.

OBJECTIVE: To investigate changes in CT manifestations and results of reverse transcription polymerase chain reaction (RT-PCR) testing between afferent and second-generation coronavirus disease 2019 (COVID-19) outside the original city (Wuhan) until recovery. METHODS: We collected 26 consecutive COVID-19 patients undergoing initial and follow-up CT scans together with RT-PCR until recovery from 2 hospitals outside the original city. Seventeen patients with afferent infection and 9 with second-generation infection were assigned to Group A and B, respectively. By observing CT manifestations, we scored COVID-19, and statistically analyzed numbers of patients with changes in CT scores and RT-PCR results between stages. RESULTS: The total score of COVID-19 on initial CT manifestations was higher in Group A than in Group B (P < 0.05). COVID-19 progressed more frequently from stage 1-2, and relieved from stage 3-4 in Group A (P < 0.05). The similar trend in Group A could not be found in Group B. Results of RT-PCR in most of patients in Group A turned negative at stage 4 while those in Group B turned negative at stage 3 (P < 0.05). CONCLUSION: Changes in CT manifestation and RT-PCR result can be different between afferent and second-generation COVID-19 until recovery.

Liver lobe based intravoxel incoherent motion diffusion weighted imaging in hepatitis B related cirrhosis: Association with child-pugh class and esophageal and gastric fundic varices.

Liver cirrhosis is a common chronic progressive liver disease in clinical practice, and intravoxel incoherent motion (IVIM) is a promising magnetic resonance method to assess liver cirrhosis, so our purpose was to investigate association of liver-lobe-based IVIM-derived parameters with hepatitis-B-related cirrhosis and its severity, and esophageal and gastric fundic varices. Seventy-four patients with hepatitis-B-related cirrhotic and 25 healthy volunteers were enrolled and underwent upper abdominal IVIM diffusion-weighted imaging with b-values of 0, 20, 50, 80, 100, 200, 400, 600, and 800 s/mm. IVIM-derived parameters (D, pure molecular diffusion; D, pseudo diffusion; and f, perfusion fraction) of left lateral lobe (LLL), left medial lobe (LML), right lobe (RL), and caudate lobe (CL) were assessed statistically to show their associations with cirrhosis and its severity, and esophageal and gastric fundic varices. In this research, we found that D, D, and f values of LLL, LML, RL, and CL were lower in cirrhotic liver than in normal liver (all P-values <.05). D, D, and f values of LLL, LML, RL, and CL were inversely correlated with Child-Pugh class of cirrhosis (r = -0.236 to -0.606, all P-values <.05). D of each liver lobe, D of LLL and CL, and f of LLL, LML, and CL in patients with esophageal and gastric fundic varices were lower than without the varices (all P-values <.05). D values of RL and CL could best identify cirrhosis, and identify esophageal and gastric fundic varices with areas under receiver-operating characteristic curve of 0.857 and 0.746, respectively. We concluded that liver-lobe-based IVIM-derived parameters can be associated with cirrhosis, and esophageal and gastric fundic varices.

CT radiomic features for predicting resectability of oesophageal squamous cell carcinoma as given by feature analysis: a case control study.

BACKGROUND: Computed tomography (CT) is commonly used in all stages of oesophageal squamous cell carcinoma (SCC) management. Compared to basic CT features, CT radiomic features can objectively obtain more information about intratumour heterogeneity. Although CT radiomics has been proved useful for predicting treatment response to chemoradiotherapy in oesophageal cancer, the best way to use CT radiomic biomarkers as predictive markers for determining resectability of oesophageal SCC remains to be developed. This study aimed to develop CT radiomic features related to resectability of oesophageal SCC with five predictive models and to determine the most predictive model. METHODS: Five hundred ninety-one patients with oesophageal SCC undergoing contrast-enhanced CT were enrolled in this study, and were composed by 270 resectable cases and 321 unresectable cases. Of the 270 resectable oesophageal SCCs, 91 cases were primary resectable tumours; and the remained 179 cases received neoadjuvant therapy after CT, shrank on therapy, and changed to resectable tumours. Four hundred thirteen oesophageal SCCs including 189 resectable cancers and 224 unresectable cancers were randomly allocated to the training cohort; and 178 oesophageal SCCs including 81 resectable tumours and 97 unresectable tumours were allocated to the validation group. Four hundred ninety-five radiomic features were extracted from CT data for identifying resectability of oesophageal SCC. Useful radiomic features were generated by dimension reduction using least absolute shrinkage and selection operator. The optimal radiomic features were chosen using multivariable logistic regression, random forest, support vector machine, X-Gradient boost and decision tree classifiers. Discriminating performance was assessed with area under receiver operating characteristic curve (AUC), accuracy and F-1score. RESULTS: Eight radiomic features were selected to create radiomic models related to resectability of oesophageal SCC (P-values < 0.01 for both cohorts). Multivariable logistic regression model showed the best performance (AUC = 0.92 ± 0.04 and 0.87 ± 0.02, accuracy = 0.87 and 0.86, and F-1score = 0.93 and 0.86 in training and validation cohorts, respectively) in comparison with any other model (P-value < 0.001). Good calibration was observed for multivariable logistic regression model. CONCLUSION: CT radiomic models could help predict resectability of oesophageal SCC, and multivariable logistic regression model is the most predictive model.

Prothrombin complex concentrate (Beriplex P/N)-related renal and cerebral infarctions in a patient with warfarin-associated intracerebral hemorrhage.

A 45-year-old man receiving warfarin treatment suffered from an intracerebral hemorrhage. Four-factor prothrombin complex concentrate (PCC) was administered to correct coagulopathy. However, bilateral renal infarcts and a cerebral infarct developed on day 5 and 7, respectively after PCC administration. Although the occurrence of PCC-related thromboembolism is low, health care practitioners should closely follow-up the symptoms and signs of thrombosis after PCC administration.