Pretreatment Magnetic Resonance Imaging Research Articles

Predictability of the radiological response to Yttrium-90 transarterial radioembolization by dynamic magnetic resonance imaging-based radiomics analysis in patients with intrahepatic cholangiocarcinoma

The study aims to investigate the predictability of the radiological response in intrahepatic cholangiocarcinoma (iCC) patients undergoing Yttrium-90 transarterial radioembolization (TARE) with a combined model built on dynamic magnetic resonance imaging (MRI)-based radiomics and clinical features. Thirty-six naive iCC patients who underwent TARE were included in this study. The tumor segmentation was performed on the axial T2-weighted (T2W) without fat suppression, axial T2W with fat suppression, and axial T1-weighted (T1W) contrast-enhanced (CE) sequence in equilibrium phase (Eq). At the sixth month MRI follow-up, all patients were divided into responders and non-responders according to the modified Response Evaluation Criteria in Solid Tumors. Subsequently, a radiomics score (rad-score) and a combined model of the rad-score and clinical features for each sequence were generated and compared between the groups. Thirteen (36.1%) patients were considered responders, and the remaining 23 (63.9%) were non-responders. Responders exhibited significantly lower rad-scores than non-responders (P < 0.050 for all sequences). The radiomics models showed good discriminatory ability with an area under the curve (AUC) of 0.696 [95% confidence interval (CI), 0.522–0.870] for the axial T1W-CE-Eq, AUC of 0.839 (95% CI, 0.709–0.970) for the axial T2W with fat suppression, and AUC of 0.836 (95% CI, 0.678–0.995) for the axial T2W without fat suppression. Radiomics models created by pre-treatment MRIs can predict the radiological response to Yttrium- 90 TARE in iCC patients with high accuracy. Combining radiomics with clinical features could increase the power of the test. Large-scale studies of multi-parametric MRIs with internal and external validations are needed to determine the clinical value of radiomics in iCC patients.

Development and validation of an MRI-Based nomogram to predict the effectiveness of immunotherapy for brain metastasis in patients with non-small cell lung cancer.

The variability and unpredictability of immune checkpoint inhibitors (ICIs) in treating brain metastases (BMs) in patients with advanced non-small cell lung cancer (NSCLC) is the main concern. We assessed the utility of novel imaging biomarkers (radiomics) for discerning patients with NSCLC and BMs who would derive advantages from ICIs treatment. Data clinical outcomes and pretreatment magnetic resonance images (MRI) were collected on patients with NSCLC with BMs treated with ICIs between June 2019 and June 2022 and divided into training and test sets. Metastatic brain lesions were contoured using ITK-SNAP software, and 3748 radiomic features capturing both intra- and peritumoral texture patterns were extracted. A clinical radiomic nomogram (CRN) was built to evaluate intracranial progression-free survival, progression-free survival, and overall survival. The prognostic value of the CRN was assessed by Kaplan-Meier survival analysis and log-rank tests. In the study, a total of 174 patients were included, and 122 and 52 were allocated to the training and validation sets correspondingly. The intratumoral radiomic signature, peritumoral radiomic signature, clinical signature, and CRN predicted intracranial objective response rate. Kaplan-Meier analyses showed a significantly longer intracranial progression-free survival in the low-CRN group than in the high-CRN group (p < 0.001). The CRN was also significantly associated with progression-free survival (p < 0.001) but not overall survival. Radiomics biomarkers from pretreatment MRI images were predictive of intracranial response. Pretreatment radiomics may allow the early prediction of benefits.

Role of multi-parametric magnetic resonance imaging in preoperative staging of cervical carcinoma in females

BackgroundIn cervical cancer, accurate diagnosis of parametrial invasion (PMI) helps in determining the next step in management, either surgical resection or neoadjuvant therapy. Cervical cancer staging is still determined according to the clinical findings. But it is widely recognized that there are differences between clinical evaluation and the actual disease extent, particularly for PMI. In this prospective study, we investigate the relationship between clinical staging and magnetic resonance imaging (MRI) staging with post-management findings among patients with cervical cancer who underwent pre-treatment MRI, using fused T2-weighted image/diffusion-weighted imaging (T2/DWI).ResultsBased on examination under anesthesia (EUA), 80% of patients showed PMI, while fused T2/DWI was positive for PMI in 77.5% of patients. In all analyzed cases (40 females), MRI had sensitivity, specificity, and accuracy of 91%, 75%, and 87.5% compared to EUA. In the operated patients (24 females), MRI likewise had sensitivity, specificity, and accuracy of 81.3%, 87.5%, and 83%, respectively. In comparison with the postoperative results, the EUA showed higher sensitivity and accuracy than MRI (92.9% vs. 81.3%) and (83.3% vs. 79.2%), respectively, in the detection of PMI in operated cases. However, MRI was more specific in the PMI detection in cases with cervical cancer than EUA (75% vs. 70%).ConclusionMRI (with the aid of fused T2/DWI) improves the diagnostic performance for the PMI assessment in cases with cervical carcinoma. Therefore, MRI is helpful in choosing surgical candidates.

Assessment of perineural spread in advanced cutaneous squamous cell carcinomas treated with immunotherapy

BackgroundCutaneous squamous cell carcinoma (CSCC) has a propensity for perineural spread (PNS) which is associated with poorer treatment outcomes. Immunotherapy is the new standard of care treatment for advanced CSCC resulting in durable responses. PNS is not captured by traditional response assessment criteria used in clinical trials, e.g. RECIST 1.1, and there is limited literature documenting radiological PNS responses to immunotherapy. In this study we assess PNS responses to immunotherapy using a modified grading system.MethodsThis is an Australian single-center retrospective review of patients with advanced CSCC who were treated with immunotherapy between April 2018 and February 2022 who had evidence of PNS on pre-treatment magnetic-resonance imaging (MRI). The primary outcome was blinded overall radiological response in PNS using graded radiological criteria, post-commencement of immunotherapy. Three defined timepoints (< 5 months, 5–10 months, > 10 months) were reviewed. Secondary outcomes included a correlation between RECIST 1.1 and PNS assessments and the assessment of PNS on fluorodeoxyglucose (FDG)-positron emission tomography (PET)/computed tomography (CT).ResultsTwenty CSCC patients treated with immunotherapy were identified. Median age was 75.7 years and 75% (n = 15) were male. All patients had locoregionally advanced disease and no distant metastases. Median follow-up was 18.5 months (range: 2–59). 70% (n = 14) demonstrated a PNS response by 5 months. Three patients experienced pseudoprogression. One patient had PNS progression by the end of study follow up. RECIST 1.1 and PNS responses were largely concordant at > 10 months (Cohen’s Kappa 0.62). 5/14 cases had features suspicious for PNS on FDG-PET/CT.ConclusionsPNS response to immunotherapy can be documented on MRI using graded radiological criteria. High response rates were seen in PNS with the use of immunotherapy in this cohort and these responses were largely concordant with RECIST 1.1 assessments. FDG-PET/CT demonstrated limited sensitivity in the detection of PNS.

Is radiomics a useful addition to magnetic resonance imaging in the preoperative classification of PitNETs?

The WHO 2021 introduced the term pituitary neuroendocrine tumours (PitNETs) for pituitary adenomas and incorporated transcription factors for subtyping, prompting the need for fresh diagnostic methods. Current biomarkers struggle to distinguish between high- and low-risk non-functioning PitNETs. We explored if radiomics can enhance preoperative decision-making. Pre-treatment magnetic resonance (MR) images of patients who underwent surgery between 2015 and 2019 with available WHO 2021 classification were used. The tumours were manually segmented on the T1w, T1-contrast enhanced, and T2w images using 3D Slicer. One hundred Pyradiomic features were extracted from each MR sequence. Models were built to classify (1) somatotroph and gonadotroph PitNETs and (2) high- and low-risk subtypes of non-functioning PitNETs. Feature were selected independently from the MR sequences and multi-sequence(combining data from more than one MRsequence) using Boruta and Pearson correlation. Support vector machine (SVM), logistic regression (LR), random forest (RF), and multi-layer perceptron (MLP) were the classifiers used. Data imbalance was addressed using the Synthetic Minority Oversampling TEchnique (SMOTE). Performance of the models were evaluated using area under the receiver operating curve (AUC), accuracy, sensitivity, and specificity. A total of 222 PitNET patients (train, n = 149; test, n = 73) were enrolled in this retrospective study. Multi-sequence-based LR model discriminated best between somatotroph and gonadotroph PitNETs, with a test AUC of 0.84, accuracy of 0.74, specificity of 0.81, and sensitivity of 0.70. Multi-sequence-based MLP model perfomed best for the high- and low-risk non-functioning PitNETs, achieving a test AUC of 0.76, accuracy of 0.67, specificity of 0.72, and sensitivity of 0.66. Utilizing pre-treatment MRI and radiomics holds promise for distinguishing high-risk from low-risk non-functioning PitNETs based on the latest WHO classification. This could assist neurosurgeons in making critical decisions regarding surgery or alternative management strategies for PitNETs after further clinical validation.

Machine learning-based MRI radiomics for assessing the level of tumor infiltrating lymphocytes in oral tongue squamous cell carcinoma: a pilot study

BackgroundTo investigate the value of machine learning (ML)-based magnetic resonance imaging (MRI) radiomics in assessing tumor-infiltrating lymphocyte (TIL) levels in patients with oral tongue squamous cell carcinoma (OTSCC).MethodsThe study included 68 patients with pathologically diagnosed OTSCC (30 with high TILs and 38 with low TILs) who underwent pretreatment MRI. Based on the regions of interest encompassing the entire tumor, a total of 750 radiomics features were extracted from T2-weighted (T2WI) and contrast-enhanced T1-weighted (ceT1WI) imaging. To reduce dimensionality, reproducibility analysis by two radiologists and collinearity analysis were performed. The top six features were selected from each sequence alone, as well as their combination, using the minimum-redundancy maximum-relevance algorithm. Random forest, logistic regression, and support vector machine models were used to predict TIL levels in OTSCC, and 10-fold cross-validation was employed to assess the performance of the classifiers.ResultsBased on the features selected from each sequence alone, the ceT1WI models outperformed the T2WI models, with a maximum area under the curve (AUC) of 0.820 versus 0.754. When combining the two sequences, the optimal features consisted of one T2WI and five ceT1WI features, all of which exhibited significant differences between patients with low and high TILs (all P < 0.05). The logistic regression model constructed using these features demonstrated the best predictive performance, with an AUC of 0.846 and an accuracy of 80.9%.ConclusionsML-based T2WI and ceT1WI radiomics can serve as valuable tools for determining the level of TILs in patients with OTSCC.

Clinical features of neurotoxicity after CD19 CAR T-cell therapy in mantle cell lymphoma

AbstractCD19 chimeric antigen receptor (CAR) T-cell therapy has proven highly effective for treating relapsed/refractory mantle cell lymphoma (MCL). However, immune effector cell–associated neurotoxicity syndrome (ICANS) remains a significant concern. This study aimed to evaluate the clinical, radiological, and laboratory correlatives associated with ICANS development after CD19 CAR T-cell therapy in patients with MCL. All patients (N = 26) who received standard-of-care brexucabtagene autoleucel until July 2022 at our institution were evaluated. Laboratory and radiographic correlatives including brain magnetic resonance imaging (MRI) and electroencephalogram (EEG) were evaluated to determine the clinical impact of ICANS. Seventeen (65%) patients experienced ICANS after treatment, with a median onset on day 6. Ten (38%) patients experienced severe (grade ≥3) ICANS. All patients with ICANS had antecedent cytokine release syndrome (CRS), but no correlation was observed between ICANS severity and CRS grade. Overall, 92% of EEGs revealed interictal changes; no patients experienced frank seizures because of ICANS. In total, 86% of patients with severe ICANS with postinfusion brain MRIs demonstrated acute neuroimaging findings not seen on pretreatment MRI. Severe ICANS was also associated with higher rates of cytopenia, coagulopathy, increased cumulative steroid exposure, and prolonged hospitalization. However, severe ICANS did not affect treatment outcomes of patients with MCL. Severe ICANS is frequently associated with a range of postinfusion brain MRI changes and abnormal EEG findings. Longer hospitalization was observed in patients with severe ICANS, especially those with abnormal acute MRI or EEG findings, but there was no discernible impact on overall treatment response and survival.

Performance and Dimensionality of Pretreatment MRI Radiomics in Rectal Carcinoma Chemoradiotherapy Prediction.

(1) Background: This study aimed to develop a machine learning model based on radiomics of pretreatment magnetic resonance imaging (MRI) 3D T2W contrast sequence scans combined with clinical parameters (CP) to predict neoadjuvant chemoradiotherapy (nCRT) response in patients with locally advanced rectal carcinoma (LARC). The study also assessed the impact of radiomics dimensionality on predictive performance. (2) Methods: Seventy-five patients were prospectively enrolled with clinicopathologically confirmed LARC and nCRT before surgery. Tumor properties were assessed by calculating 2141 radiomics features. Least absolute shrinkage selection operator (LASSO) and multivariate regression were used for feature selection. (3) Results: Two predictive models were constructed, one starting from 72 CP and 107 radiomics features, and the other from 72 CP and 1862 radiomics features. The models revealed moderately advantageous impact of increased dimensionality, with their predictive respective AUCs of 0.86 and 0.90 in the entire cohort and 0.84 within validation folds. Both models outperformed the CP-only model (AUC = 0.80) which served as the benchmark for predictive performance without radiomics. (4) Conclusions: Predictive models developed in this study combining pretreatment MRI radiomics and clinicopathological features may potentially provide a routine clinical predictor of chemoradiotherapy responders, enabling clinicians to personalize treatment strategies for rectal carcinoma.

Development of an MRI-Based Prediction Model for Anti-TNF Treatment Failure in Perianal Crohn’s Disease: A Multicenter Study

Clinical and radiologic variables associated with perianal fistula (PAF) outcomes are poorly understood. We developed prediction models for anti-tumor necrosis factor (TNF) treatment failure in patients with Crohn's disease-related PAF. In a multicenter retrospective study between 2005 and 2022 we included biologic-naive adults (>17 years) who initiated their first anti-TNF therapy for PAF after pelvic magnetic resonance imaging (MRI). Pretreatment MRI studies were prospectively reread centrally by blinded radiologists. We developed and internally validated a prediction model based on clinical and radiologic parameters to predict the likelihood of anti-TNF treatment failure, clinically, at 6 months. We compared our model and a simplified version of MRI parameters alone with existing imaging-based PAF activity indices (MAGNIFI-CD and modified Van Assche MRI scores) by De Long statistical test. We included 221 patients: 32 ± 14 years, 60% males, 76% complex fistulas; 68% treated with infliximab and 32% treated with adalimumab. Treatment failure occurred in 102 (46%) patients. Our prediction model included age at PAF diagnosis, time to initiate anti-TNF treatment, and smoking and 8 MRI characteristics (supra/extrasphincteric anatomy, fistula length >4.3 cm, primary tracts >1, secondary tracts >1, external openings >1, tract hyperintensity on T1-weighted imaging, horseshoe anatomy, and collections >1.3 cm). Our full and simplified MRI models had fair discriminatory capacity for anti-TNF treatment failure (concordance statistic, 0.67 and 0.65, respectively) and outperformed MAGNIFI-CD (P= .002 and < .0005) and modified Van Assche MRI scores (P < .0001 and < .0001), respectively. Our risk prediction models consisting of clinical and/or radiologic variables accurately predict treatment failure in patients with PAF.

Prognostic value of the ratio of pretreatment carcinoembryonic antigen to tumor volume in rectal cancer.

As a commonly used biomarker in rectal cancer (RC), the prognostic value of carcinoembryonic antigen (CEA) remains underexplored. This study aims to evaluate the prognostic value of pretreatment CEA/tumor volume in RC. This retrospective study included patients who underwent pretreatment magnetic resonance imaging (MRI) with histologically confirmed primary rectal adenocarcinoma from November 2012 to April 2018. Patients were divided into high-risk and low-risk groups according to the median values of CEA/Diapath (CEA to pathological diameter), CEA/DiaMRI (CEA to MRI tumor diameter), and CEA/VolMRI (CEA to MRI tumor volume). Cox regression analysis was utilized to determine the prognostic value of CEA, CEA/Diapath, CEA/DiaMRI, and CEA/VolMRI. Stepwise regression was used to establish nomograms for predicting disease-free survival (DFS) and overall survival (OS). Predictive performance was estimated by using the concordance index (C-index) and area under curve receiver operating characteristic (AUC). A total of 343 patients [median age 58.99 years, 206 (60.06%) males] were included. After adjusting for patient-related and tumor-related factors, CEA/VolMRI was superior to CEA, CEA/Diapath, and CEA/DiaMRI in distinguishing high-risk from low-risk patients in terms of DFS [hazard ratio (HR) =1.83; P=0.010] and OS (HR =1.67; P=0.048). Subanalysis revealed that CEA/VolMRI stratified high death risk in CEA-negative individuals (HR =2.50; P=0.038), and also stratified low recurrence risk in CEA-positive individuals (HR =2.06; P=0.024). In the subanalysis of stage II or III cases, the highest HRs and the smallest P values were observed in distinguishing high-risk from low-risk patients according to CEA/VolMRI in terms of DFS (HR =2.44; P=0.046 or HR =2.41; P=0.001) and OS (HR =1.96; P=0.130 or HR =2.22; P=0.008). The nomograms incorporating CEA/VolMRI showed good performance, with a C-index of 0.72 [95% confidence interval (CI): 0.68-0.79] for DFS and 0.73 (95% CI: 0.68-0.80) for OS. Higher CEA/VolMRI was associated with worse DFS and OS. CEA/VolMRI was superior to CEA, CEA/Diapath, and CEA/DiaMRI in predicting DFS and OS. Pretreatment CEA/VolMRI may facilitate risk stratification and treatment decision-making.

Imaging-proteomic analysis for prediction of neoadjuvant chemotherapy responses in patients with breast cancer.

Optimizing patient selection for neoadjuvant chemotherapy in patients with breast cancer remains an unmet clinical need. Quantitative features from medical imaging were reported to be predictive of treatment responses. However, the biologic meaning of these latent features is poorly understood, preventing the clinical use of such noninvasive imaging markers. The study aimed to develop a deep learning signature (DLS) from pretreatment magnetic resonance imaging (MRI) for predicting responses to neoadjuvant chemotherapy in patients with breast cancer and to further investigate the biologic meaning of the DLS by identifying its underlying pathways using paired MRI and proteomic sequencing data. MRI-based DLS was constructed (radiogenomic training dataset, n = 105) and validated (radiogenomic validation dataset, n = 26) for the prediction of pathologic complete response (pCR) to neoadjuvant chemotherapy. Proteomic sequencing revealed biological functions facilitating pCR (n = 139). Their associations with DLS were uncovered by radiogenomic analysis. The DLS achieved a prediction accuracy of 0.923 with an AUC of 0.958, higher than the performance of the model trained by transfer learning. Cellular membrane formation, endocytosis, insulin-like growth factor binding, protein localization to membranes, and cytoskeleton-dependent trafficking were differentially regulated in patients showing pCR. Oncogenic signaling pathways, features correlated with human phenotypes, and features correlated with general biological processes were significantly correlated with DLS in both training and validation dataset (p.adj < 0.05). Our study offers a biologically interpretable DLS for the prediction of pCR to neoadjuvant chemotherapy in patients with breast cancer, which may guide personalized medication.

Patient survival prediction in locally advanced cervical squamous cell carcinoma using MRI-based radiomics: retrospective cohort study

Cervical cancer is a major health concern for women, ranking as the fourth most common cancer and a significant cause of cancer-related deaths worldwide. To enhance prognostic predictions for locally advanced cervical squamous cell carcinoma, we conducted a study utilizing radiomics features extracted from pretreatment magnetic resonance images. The goal was to predict patient survival and compare the predictive value of these features with clinical traits and the 2018 International Federation of Obstetrics and Gynecology (FIGO) staging system. In our retrospective cohort study, we included 500 patients with confirmed cervical squamous cell carcinoma ranging from FIGO stages IIB to IVA under the 2018 staging system. All patients underwent pelvic MRI with diffusion-weighted imaging before receiving definitive curative concurrent chemoradiotherapy. The results showed that the combination model, incorporating radiomics scores and clinical traits, demonstrated superior predictive accuracy compared to the widely used 2018 FIGO staging system for both progression-free and overall survival. Age was identified as a significant factor influencing survival outcomes. Additionally, primary tumour invasion stage, tumour maximal diameter, and the location of lymph node metastasis were found to be important predictors of progression-free survival, while primary tumour invasion stage and lymph node metastasis position individually affected overall survival. During the follow-up period, a portion of patients experienced disease-related deaths or tumour progression/recurrence in both sets. The radiomics-score significantly enhanced prediction ability, providing valuable insights for guiding personalized therapy approaches and stratifying patients into low-risk and high-risk categories for progression-free and overall survival. In conclusion, our study demonstrated the potential of radiomics features as a valuable addition to existing clinical tools like the FIGO staging system, offering promising advancements in managing locally advanced cervical squamous cell carcinoma.

Using MRI Radiomics to Predict the Efficacy of Immunotherapy for Brain Metastasis in Patients with Small Cell Lung Cancer

Immune-checkpoint inhibitors (ICIs) combined with chemotherapy has been widely used in the first-line treatment of small cell lung cancer (SCLC) patients. However, the efficacy of ICIs for patients with brain metastases (BMs) of SCLC is limited. There are no effective factors to predict the efficacy. We developed and validate a model to predict intracranial efficacy for ICIs in patients with BMs from SCLC by using MRI radiomics. In this study, we collected 101 SCLC patients with BMs treated with ICIs. They clinical characteristics and pre-treatment Magnetic Resonance Imaging (MRI) were collected. Seventy cases collected from our hospital as training cohort and 31 cases collected from another hospital as an independent validation cohort. Brain metastatic lesions were contoured on ITK-SNAP software and 3748 radiomic features capturing both intra- and peritumoral texture patterns were extracted. The primary endpoint of this study was intracranial overall response rate (ORR). Intraclass correlation coefficient (ICC) and random forest (RF) model were used to select radiomic features. The top 10 selected radiomic features were adopted to build prognostic models by using logistic regression. Area Under the Curve (AUC) of the Receiver Operating Characteristic (ROC) was chosen as the metric to assess model performances. We also performed a nomogram based on a multivariate logistic regression model includes radiomic and clinical features to predict ICIs intracranial efficacy. The clinical characteristics including number of treatment lines and concurrent brain radiotherapy, ten intratumoral signatures and one peritumoral signatures were found significantly associated with ICIs intracranial efficacy. Predictors contained in the individualized prediction nomogram included the radiomics signature, and clinical features. The model showed favorable discrimination with a C-index of 0.861[95% CI: 0.779-0.943] and AUC of 0.861[95% CI: 0.778- 0.944] in training cohort and with a C-index of 0.812[95% CI: 0.647-0.978] and AUC of 0.812[95% CI:0.646-0.979] in the test cohort. Radiomic features from pre-treatment MRI images were predictive for intracranial efficacy in SCLC patients with BMs. Pre-treatment radiomics may allow early prediction of benefit and expedite more aggressive treatment for high-risk patients and it has to be further explored as predictor of outcome in a larger series of patients.

Prediction of Progression After Cervix Cancer Radiotherapy Using a Machine-Learning Model on Pre-Treatment MRI

MRI may be useful to identify women with cervical cancer at high risk of disease progression to test strategies of treatment intensification. The purpose of this study was to determine the value of a machine-learning model built on pre-treatment MRI for prediction of risk of progression after radiation therapy. MagneticResonance Imaging (MRI) data for women with cervical cancer was collected from The Cancer Genome Atlas Cervical Squamous Cell Carcinoma and Endocervical Adenocarcinoma Collection (TCGA-CESC) on the Cancer Imaging Archive (TCIA), which reported clinical, treatment, and imaging data from a single institution. 27 patients who had received radiation for cervical cancer were selected for input into a custom 3-D Residual Neural Network (ResNet) model with added custom layers specific to DICOM data in tensorflow python package. One T2 MRI per patient was used to predict recurrence free survival after radiation treatment, where patients were predicted to be "high risk" or "low risk" for disease recurrence as the output of the model. All slices of the T2 MRI were used. The model was validated using five-fold cross validation; 80% of the data was used to train each fold and 20% was used for testing. Final model statistical significance was confirmed through shuffle test at the p < 0.01 level. The clinical outcomes of patients and the model's "low-risk" and "high-risk" prediction were compared. There were 27 patients in the study with mean age of 51 years (range 29-79). 20 patients had squamous cell carcinoma and 7 patients had adenocarcinoma. The stage breakdown consisted of 9 women IB, 2 IIA, 9 IIB, 2 IIIA, 2 IIIB, and 3 stage IV. 10 women were treated with radiation alone and 17 with chemo-radiation. 5 women received surgery in addition to radiation or chemoradiation. 21 patients received brachytherapy. Median follow-up of patients was 29 months (range 3-64). The model predicted 7 patients as "high risk" for recurrence; all 7 developed a recurrence during follow up. None of the 20 patients predicted to be "low risk" developed disease recurrence. Among all patients in the study, the two-year progression free survival (PFS) was 82.0%. Patients identified as "low risk" and "high risk" by model had two-year PFS of 100% and 43%, respectively. Among patients with recurrence, 3 developed local recurrence and 4 developed distant metastases. The ResNet model achieved cross-validated accuracy of 92% for prediction of progression-free survival (p<0.01). A 3-D ResNet machine-learning model using pretreatment MRI image data can accurately predict clinical outcomes for cervical cancer following radiation therapy. Future work to confirm generalizability should focus on validation with a larger clinical dataset.

Predicting the BRAF mutation with pretreatment MRI radiomics features for melanoma brain metastases receiving Gamma Knife radiosurgery

To develop a model using radiomics features extracted from magnetic resonance imaging (MRI) images of Gamma Knife radiosurgery (GKRS) to predict the BRAF mutation in patients with melanoma brain metastases (MBM). Data of 220 tumours were classified into two groups. One was a group whose BRAF mutation was identified, and the other group whose BRAF mutation was not identified. We extracted 1,962 radiomics features from gadolinium contrast-enhanced T1-weighted MRI treatment-planning images. Synthetic Minority Over-sampling TEchnique (SMOTE) was performed to address the unbalanced data-related issues. A single-layer neural network (NN) was used to build predictive models with radiomics features. The sensitivity, specificity, accuracy, and the area under the curve (AUC) were evaluated to assess the model performance. The prediction performance for the final evaluation without the SMOTE had an accuracy of 77.14%, a specificity of 82.44%, a sensitivity of 81.85%, and an AUC of 0.79. The application of SMOTE improved the prediction model to an accuracy of 83.1%, a specificity of 87.07%, a sensitivity of 78.82%, and an AUC of 0.82. The current study showed the feasibility of generating a highly accurate NN model for the BRAF mutation prediction. The prediction performance improved with SMOTE. The model assists physicians to obtain more accurate expectations of the treatment outcome without a genetic test.

Noninvasive prediction of insufficient biochemical response after ursodeoxycholic acid treatment in patients with primary biliary cholangitis based on pretreatment nonenhanced MRI

ObjectivesTo explore the feasibility of pretreatment nonenhanced magnetic resonance imaging (MRI) in predicting insufficient biochemical response to ursodeoxycholic acid (UDCA) in patients with primary biliary cholangitis (PBC).MethodsFrom January 2009 to April 2022, consecutive PBC patients who were treated with UDCA and underwent nonenhanced MRI within 30 days before treatment were retrospectively enrolled. All MR images were independently evaluated by two blinded radiologists. Uni- and multivariable logistic regression analyses were performed to develop a predictive model for 12-month insufficient biochemical response. Model performances were evaluated by computing the area under the receiver operating characteristic curve (AUC), sensitivity, and specificity.ResultsA total of 74 patients (50.6 ± 11.9 years; 62 females) were included. Three pretreatment MRI features, including hepatomegaly (odds ratio [OR]: 4.580; p = 0.011), periportal hyperintensity on T2-weighted imaging (T2WI) (OR: 4.795, p = 0.008), and narrowing of the bile ducts (OR: 3.491; p = 0.027) were associated with 12-month insufficient biochemical response in the multivariable analysis. A predictive model based on the above indicators had an AUC of 0.781, sensitivity of 85.4%, and specificity of 61.5% for predicting insufficient biochemical response.ConclusionsA noninvasive model based on three pretreatment MRI features could accurately predict 12-month insufficient biochemical response to UDCA in patients with PBC. Early identification of PBC patients at increased risk for insufficient response can facilitate the timely initiation of additional treatment.Clinical relevance statementA noninvasive predictive model constructed by incorporating three pretreatment MRI features may help identify patients with primary biliary cholangitis at high risk of insufficient biochemical response to ursodeoxycholic acid and facilitate the timely initiation of additional treatment.Key Points• Noninvasive imaging features based on nonenhanced pretreatment MRI may predict an insufficient biochemical response to UDCA in PBC patients.• A combined model based on three MRI features (hepatomegaly, periportal hyperintensity on T2-weighted imaging, and narrowing of the bile ducts) further improved the predictive efficacy for an insufficient biochemical response to UDCA in PBC patients, with high sensitivity and specificity.• The nomogram of the combined model showed good calibration and predictive efficacy for an insufficient biochemical response to UDCA in PBC patients. In particular, the calibration curve visualised the clinical applicability of the prediction model.

Incidence and Predictors of Relapse After Weaning Immune Suppressive Therapy in Cardiac Sarcoidosis

Cardiac sarcoidosis (CS) is a relapsing-remitting disease, and immune suppression (IS) is the mainstay of therapy. Predictors of relapse for patients with CS in remission are not well characterized. We assessed incidence of relapse in consecutive patients with CS treated with high-dose steroids and/or steroid-sparing agents (SSA) in our center from 2000 to 2020. Remission was defined as reaching maintenance therapy (no IS, SSA, and/or prednisone ≤5mg/d) for ≥1month. Relapse was defined as recurrence of CS syndrome requiring IS intensification: heart failure, ventricular arrhythmia, decrease in left ventricular ejection fraction, or increased disease burden on imaging. Among 68 patients, the mean age was 50.7±9.0years; 25 (37%) were women, and 32 (47%) were Black. In total, 59 patients (87%) reached remission. Over a median follow-up of 39.5months (interquartile range 17.6, 92.5), 28 (48%) relapsed. Greater percentage of late gadolinium enhancement (LGE) on pretreatment magnetic resonance imaging corresponded with increased likelihood of relapse (odds ratio 1.396 per 5% increase [95% confidence interval (CI) 1.04 to 1.88]; p=0.028). LGE ≥11% predicted elevated risk of relapse (adjusted odds ratio 4.998 [1.34 to 18.64]; p=0.017). Shorter time to relapse was observed with isolated CS (adjusted hazard ratio 4.084 [1.44,11.56]; p=0.008) and LGE ≥11% (adjusted hazard ratio 3.007 [1.01, 8.98]; p=0.049). Approximately 1 in 2 patients with CS in remission experienced relapse. Greater burden of LGE on cardiac magnetic resonance imaging and isolated CS are associated with greater risk of relapse. Future work is needed to refine risk stratification for relapse and to optimize surveillance strategies on the basis of the burden of disease.

MRI features for prediction of the intravenous chemotherapy effect in patients with retinoblastoma

To investigate the value of orbital magnetic resonance imaging (MRI) features in predicting the efficacy of intravenous chemotherapy (IVC) for patients with retinoblastoma (RB). The pretreatment clinical and MRI data of 100 eyes from 80 RB patients who underwent IVC were collected retrospectively. There were 59 eyes in the effective group and 41 eyes in the ineffective group, and the baseline data of the two groups were compared statistically. Three radiologists reviewed and evaluated each lesion independently based on 25 MRI features. The predictive values of the MRI features for IVC efficacy were assessed by multi-factor logistic regression analysis, and their odds ratios (ORs) and 95% confidence intervals (95% CIs) were calculated. Receiver operating characteristic curves (ROCs) with the area under the curve (AUC) were used to determine the predictive abilities. A predictive model was constructed by integrating all independent predictors visualised by the nomogram. There were no statistically significant differences in sex or age between the effective and ineffective groups. The results of multivariate regression analysis showed that laterality, margin, and anterior eye segment enhancement were identified as independent factors that could predict IVC efficacy. The predictive model combining these three features was constructed, and it had an AUC of 0.732 (95% CI: 0.633, 0.831, p<0.01), a sensitivity of 71.2%, and a specificity of 70.7%. The data demonstrate that the orbital MRI features can be used to predict IVC efficiency before RB patients are treated.

Association Between Pseudoprogression of Vestibular Schwannoma After Radiosurgery and Radiological Features of Solid and Cystic Components.

The pathophysiology of vestibular schwannoma (VS) pseudoprogression after Gamma Knife radiosurgery (GKRS) remains unclear. Radiological features in pretreatment magnetic resonance images may help predict VS pseudoprogression. This study used VS radiological features quantified using an automated segmentation algorithm to predict pseudoprogression after GKRS treatment. This is a retrospective study comprising 330 patients with VS who received GKRS. After image preprocessing and T2W/contrast-enhanced T1-weighted image (CET1W) image generation, with fuzzy C-means clustering, VSs were segmented into solid and cystic components and classified as solid and cystic. Relevant radiological features were then extracted. The response to GKRS was classified into "nonpseudoprogression" and "pseudoprogression/fluctuation". The Z test for two proportions was used to compare solid and cystic VS for the likelihood of pseudoprogression/fluctuation. Logistic regression was used to assess the correlation between clinical variables and radiological features and response to GKRS. The likelihood of pseudoprogression/fluctuation after GKRS was significantly higher for solid VS compared with cystic VS (55% vs 31%, P < .001). For the entire VS cohort, multivariable logistic regression revealed that a lower mean tumor signal intensity (SI) in T2W/CET1W images was associated with pseudoprogression/fluctuation after GKRS ( P = .001). For the solid VS subgroup, a lower mean tumor SI in T2W/CET1W images ( P = .035) was associated with pseudoprogression/fluctuation after GKRS. For the cystic VS subgroup, a lower mean SI of the cystic component in T2W/CET1W images ( P = .040) was associated with pseudoprogression/fluctuation after GKRS. Pseudoprogression is more likely to occur in solid VS compared with cystic VS. Quantitative radiological features in pretreatment magnetic resonance images were associated with pseudoprogression after GKRS. In T2W/CET1W images, solid VS with a lower mean tumor SI and cystic VS with a lower mean SI of cystic component were more likely to have pseudoprogression after GKRS. These radiological features can help predict the likelihood of pseudoprogression after GKRS.

Poster 253: Is the Use of Pretreatment Magnetic Resonance Imaging Necessary Prior to Surgery in Pediatric Tibial Spine Fractures

Objectives: Injuries to the tibial spine occur as a result of an anterior cruciate ligament (ACL) avulsion fracture from the intercondylar eminence of the proximal tibia. Similar to anterior cruciate ligament ruptures, tibial spine injuries are associated with damage to other intraarticular structures. These associated injuries are often visualized with Magnetic Resonance Imaging (MRI) obtained prior to surgery. The purpose of the study is to describe concomitant injuries visualized by a pretreatment MRI, to examine how this correlates to what is seen at the time of surgery, and to determine how a pretreatment MRI affects the time between injury and surgery as well the length of the surgical procedure itself. Our hypothesis is that a pretreatment MRI results in a longer time between injury and surgery without any affect on operating time or amount of concomitant pathology identified. Methods: Utilizing an institutional review board approved retrospective study, we identified patients under 18 years of age who underwent arthroscopic management of a tibial spine fracture over a 13 year period. Our exclusion criteria were patients with a concomitant ipsilateral lower extremity fracture or posterior cruciate ligament injury, patients with poor imaging quality, and patients with incomplete medical records. Patients were then categorized based on whether they obtained a pretreatment MRI. We then evaluated for concomitant injuries, time from injury to surgery, total room time, and tourniquet time between the groups. Results: A total of 85 patients met this inclusion criteria, of which 44 (51.7%) had a pretreatment MRI. There were no significant differences in age, gender, or Myers and McKeever grade between groups. All were treated with arthroscopic reduction and fixation. In the MRI group, there were 30 patients (68.2%) that were identified to have additional meniscal or chondral pathology. The three most common pathologies seen on MRI were lateral meniscus tears (13 patients), medial meniscus tears (6 patients), and intermeniscal ligament entrapment (6 patients). In addition, 28 patients (63.6%) were found to have a bone contusion pattern similar to that seen in ACL injuries and 17 patients (38.6%) showed signs of damage to the ACL fibers themselves. Of those 30 patients with meniscal pathology or entrapment seen on MRI, 17 patients (56.7%) were confirmed to have that pathology at the time of arthroscopy. The percentage of patients with meniscal or chondral pathology seen on arthroscopy was not significantly different between those who received a pretreatment MRI and those who did not (38.6% MRI group vs 41.5% No-MRI group p =0.79). The time from injury to surgery was significantly longer in the MRI group (19.3 days vs 10.1 days p=0.01). Both the tourniquet time (72 min vs 76 min) and operative time (83.5 min vs 90.5 min) were not significantly different(p=0.49 for tourniquet time and p=0.32 for room time). Conclusions: Over two-thirds of patients who sustained a tibial spine fracture were noted to have concomitant pathology on MRI, with 56.7% of injuries seen on MRI correlating to what was seen at the time of surgery. The use of pretreatment MRI was shown to delay surgery without any significant change in operative time, tourniquet time, or percentage of pathology identified at time of surgery. Previous studies have suggested that a pre-treatment MRI may be helpful in identifying pathology, however we believe this is the first study to determine if obtaining a pretreatment MRI influences operative time, tourniquet time, and pathology identified at time of surgery.