Deep-Learning-Based Predictive Imaging Biomarker Model for EGFR Mutation Status in Non-Small Cell Lung Cancer from CT Imaging.

PURPOSE: The authors aimed to develop and validate deep-learning-based radiogenomic (DLR) models and radiomic signatures to predict the EGFR mutation in patients with NSCLC, and to assess the semantic and clinical features that can contribute to detecting EGFR mutations. METHODS: Using 990 patients from two NSCLC trials, we employed an end-to-end pipeline analyzing CT images without precise segmentation. Two 3D convolutional neural networks segmented lung masses and nodules. RESULTS: The combined radiomics and DLR model achieved an AUC of 0.88 ± 0.03 in predicting EGFR mutation status, outperforming individual models. Semantic features further improved the model's accuracy, with an AUC of 0.88 ± 0.05. CT semantic features that were found to be significantly associated with EGFR mutations were pure solid tumours with no associated ground glass component (p < 0.03), the absence of peripheral emphysema (p < 0.03), the presence of pleural retraction (p = 0.004), the presence of fissure attachment (p = 0.001), the presence of metastatic nodules in both the tumour-containing lobe (p = 0.001) and the non-tumour-containing lobe (p = 0.001), the presence of ipsilateral pleural effusion (p = 0.04), and average enhancement of the tumour mass above 54 HU (p < 0.001). CONCLUSIONS: This AI-based radiomics and DLR model demonstrated high accuracy in predicting EGFR mutation, serving as a non-invasive and user-friendly imaging biomarker for EGFR mutation status prediction.

Role of imaging in the management of thyroglossal duct cyst carcinomas (TGC-TIRADS): a single centre retrospective study over 16 years.

Introduction: Thyroglossal duct cyst (TGDC) is the most frequently encountered developmental anomaly in thyroid genesis with a reported incidence of 7% in the adult population. The cyst is known to develop anywhere along the pathway of thyroid descent but is more frequently seen in the infrahyoid neck in the midline. The incidence of malignancy in a TGDC is approximately 1%; a majority of these are papillary carcinomas. This study was conducted at a single tertiary care centre which spanned over a decade which adds practice changing evidence-based knowledge to existing literature on this rare entity. A comprehensive study which conclusively establishes the imaging features predictive of malignancy in TGDC carcinomas (TGDCa), the protocol for optimal management, clinical outcome and long-term survival of these patients is not available. Although TGDC carcinoma is thought to have an excellent prognosis, there is not enough data available on the long-term survival of these patients. The aim of this study was to identify whether neck ultrasound (US) can serve as an accurate imaging tool for the preoperative diagnosis of TGDC carcinomas. Methods: We accessed the electronic medical records of 86 patients with TGDC between January 2005 to December 2021. Of these, 22 patients were detected with TGDC papillary carcinoma on histopathologic examination. Relevant imaging, treatment and follow up information for all cases of TGDC carcinoma were retrospectively reviewed. We compared US characteristics predictive of malignancy across outcomes groups; malignant vs benign using the Chi-square test. Based on the results, a TGC-TIRADS classification was proposed with calculation of the percentage likelihood of malignancy for each category. Results: Compared to benign TGDCs, malignant TGDCs were more likely to present with following US characteristics: irregular or lobulated margins (90.40 vs. 38.10%), solid-cystic composition (61.90 vs. 17.07%), internal vascularity (47.62 vs. 4.88 %), internal calcification (76.19 vs. 7.32 %) (each p value < 0.005). Calcifications and internal vascularity were the most specific while irregular/lobulated margins were the most sensitive feature for malignancy. AUC under the ROC curve was 0.88. Allpatients were operated and were disease free at the end of 5 years or till the recent follow up. Discussion: US is the imaging modality of choice for pre-operative diagnosis of TGDC carcinoma. Thepre-operative diagnosis and risk stratification of thyroglossal lesions will be aided by the application of the proposed TGC-TIRADS classification, for which the percentage likelihood of malignancy correlated well with the results in our study. Sistrunk procedure is adequate for isolated TGDC carcinoma; suspicious neck nodes on imaging also necessitates selective nodal dissection. Papillary carcinomas have an excellent prognosis with low incidence of disease recurrence.

Exploring the implications of modified advanced lung cancer inflammation index on outcomes in patients with advanced non-small cell lung cancer.

Aim: Sarcopenia and skeletal muscle density (SMD) have been shown to be both predictive and prognostic marker in oncology. Advanced lung cancer inflammation index (ALI) has been shown to predict overall survival (OS) in small cell lung cancer (SCLC). Computed tomography (CT) enables skeletal muscle to be quantified, whereas body mass index (BMI) cannot accurately reflect body composition. The purpose was to evaluate the prognostic value of modified ALI (mALI) using CT-determined third lumbar vertebra (L3) muscle index beyond original ALI and see the interaction between sarcopenia, SMD, neutrophil-lymphocyte ratio (NLR), ALI and mALI at baseline and post 4 cycles of chemotherapy and their effects on OS and progress free survival (PFS) in patients with advanced non-SCLC (NSCLC). Methods: This retrospective study consisted of a total of 285 advanced NSCLC patients. The morphometric parameters such as SMD, skeletal muscle index (SMI) and fat-free mass (FFM) were measured by CT at the L3 vertebra. ALI was defined as BMI × serum albumin/NLR and mALI was defined as SMI × serum albumin/NLR. Results: Sarcopenia was observed in over 70% of patients across all BMI categories. Patients having sarcopenia suffered from a higher incidence of chemotherapeutic drug toxicities but this was not found to be statistically significant. Concordance was seen between ALI and mALI in the pre-treatment setting and this was statistically significant. A significant proportion of patients with poor ALI (90.9%), poor pre-chemotherapy mALI (91.3%) and poor post-chemotherapy mALI (89%) had poor NLR and each of them was statistically significant. Conclusions: In both univariate and multivariate analyses, this study demonstrated the statistical significance of sarcopenia, SMD, and mALI as predictive factors for OS. Additionally, sarcopenia and SMD were also found to be statistically significant factors in predicting PFS. These biomarkers could potentially help triage patients for active nutritional intervention for better outcomes.

Response assessment of post-treatment head and neck cancers to determine further management using NI-RADS (Neck Imaging Reporting and Data System): a subgroup analysis of a randomized controlled trial.

Objective: Interpreting complex post-treatment changes in head and neck cancer (HNC) is challenging with further added perplexity due to variable interobserver interpretation and hence evolved the NI-RADS lexicon. We evaluated the accuracy of NI-RADS in predicting disease status on 1st post-treatment follow-up CECT in a homogenous cohort of those who received only chemoradiation. Methods: Retrospective analysis of imaging was done for LASHNC patients who received radical chemoradiation in an open-label, investigator-initiated, phase 3 randomized trial (2012-2018) randomly assigned to either radical radiotherapy with concurrent weekly cisplatin (CRT) or CRT with the same schedule plus weekly nimotuzumab (NCRT). 536 patients were accrued, and 74 patients who did not undergo PET/CECT after 8 weeks post-CRT were excluded. After assessing 462 patients for eligibility to allocate NI-RADS at primary and node sites, 435 cases fell in the Primary disease cohort and 412 cases in the Node disease cohort. We evaluated sensitivity, disease prevalence, the positive and negative predictive value of the NI-RADS lexicon, and accuracy, which were expressed as percentages. We also prepared flow charts to determine concordance with allocated NI-RADS category and established accuracy with which it can identify disease status. Results: Out of 435 primary disease cohort, 92%, 55%, 48%,70% were concordant and had 100%, 72%, 70%, 82% accuracy in NI-RADS1 (n=12), NI-RADS2 (n=261), NIRADS3 (n=105), and NI-RADS 4 (n=60) respectively. Out of 412 nodes disease cohort, 95%, 90%, 48%, 70%were concordant and had 92%, 97%, 90%, 67% accuracy in NI-RADS1 (n=57), NI-RADS2 (n=255), NI-RADS3 (n=105) and NI-RADS4 (n=60) respectively. % concordance of PET/CT and CECT across all primary and node disease cohorts revealed that PET/CT was 91% concordant in primary NI-RADS2 as compared to 55% concordance of CECT whereas concordance of CECT was better with 57% in primary NI-RADS3 cohort as compared to PET/CT concordance of 41%. Conclusion: The accuracy with which the NI-RADS lexicon performed in our study at node sites was better than that at the primary site. There is a great scope of research to understand if CECT performs better over clinical disease status in NI-RADS 3 and 4 categories. Further research should be carried out to understand if PET/CECT can be used for close interval follow-up in stage III/IV NI-RADS 2 cases.

Deep learning based automated epidermal growth factor receptor and anaplastic lymphoma kinase status prediction of brain metastasis in non-small cell lung cancer.

Aim: The aim of this study was to investigate the feasibility of developing a deep learning (DL) algorithm for classifying brain metastases from non-small cell lung cancer (NSCLC) into epidermal growth factor receptor (EGFR) mutation and anaplastic lymphoma kinase (ALK) rearrangement groups and to compare the accuracy with classification based on semantic features on imaging. Methods: Data set of 117 patients was analysed from 2014 to 2018 out of which 33 patients were EGFR positive, 43 patients were ALK positive and 41 patients were negative for either mutation. Convolutional neural network (CNN) architecture efficient net was used to study the accuracy of classification using T1 weighted (T1W) magnetic resonance imaging (MRI) sequence, T2 weighted (T2W) MRI sequence, T1W post contrast (T1post) MRI sequence, fluid attenuated inversion recovery (FLAIR) MRI sequences. The dataset was divided into 80% training and 20% testing. The associations between mutation status and semantic features, specifically sex, smoking history, EGFR mutation and ALK rearrangement status, extracranial metastasis, performance status and imaging variables of brain metastasis were analysed using descriptive analysis [chi-square test (χ2)], univariate and multivariate logistic regression analysis assuming 95% confidence interval (CI). Results: In this study of 117 patients, the analysis by semantic method showed 79.2% of the patients belonged to ALK positive were non-smokers as compared to double negative groups (P = 0.03). There was a 10-fold increase in ALK positivity as compared to EGFR positivity in ring enhancing lesions patients (P = 0.015) and there was also a 6.4-fold increase in ALK positivity as compared to double negative groups in meningeal involvement patients (P = 0.004). Using CNN Efficient Net DL model, the study achieved 76% accuracy in classifying ALK rearrangement and EGFR mutations without manual segmentation of metastatic lesions. Analysis of the manually segmented dataset resulted in improved accuracy of 89% through this model. Conclusions: Both semantic features and DL model showed comparable accuracy in classifying EGFR mutation and ALK rearrangement. Both methods can be clinically used to predict mutation status while biopsy or genetic testing is undertaken.

Deep learning based clinico-radiological model for paediatric brain tumor detection and subtype prediction.

Aim: Early diagnosis of paediatric brain tumors significantly improves the outcome. The aim is to study magnetic resonance imaging (MRI) features of paediatric brain tumors and to develop an automated segmentation (AS) tool which could segment and classify tumors using deep learning methods and compare with radiologist assessment. Methods: This study included 94 cases, of which 75 were diagnosed cases of ependymoma, medulloblastoma, brainstem glioma, and pilocytic astrocytoma and 19 were normal MRI brain cases. The data was randomized into training data, 64 cases; test data, 21 cases and validation data, 9 cases to devise a deep learning algorithm to segment the paediatric brain tumor. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of the deep learning model were compared with radiologist's findings. Performance evaluation of AS was done based on Dice score and Hausdorff95 distance. Results: Analysis of MRI semantic features was done with necrosis and haemorrhage as predicting features for ependymoma, diffusion restriction and cystic changes were predictors for medulloblastoma. The accuracy of detecting abnormalities was 90%, with a specificity of 100%. Further segmentation of the tumor into enhancing and non-enhancing components was done. The segmentation results for whole tumor (WT), enhancing tumor (ET), and non-enhancing tumor (NET) have been analyzed by Dice score and Hausdorff95 distance. The accuracy of prediction of all MRI features was compared with experienced radiologist's findings. Substantial agreement observed between the classification by model and the radiologist's given classification [K-0.695 (K is Cohen's kappa score for interrater reliability)]. Conclusions: The deep learning model had very high accuracy and specificity for predicting the magnetic resonance (MR) characteristics and close to 80% accuracy in predicting tumor type. This model can serve as a potential tool to make a timely and accurate diagnosis for radiologists not trained in neuroradiology.

CT-Based Screening for Pulmonary Metastases in Head and Neck Squamous Cell Cancers: Diagnostic Accuracy and Cost Comparison with PET-CECT.

This study's objective was to compare detection rates of radiograph, computed tomography (CT), and positron emission tomography-contrast-enhanced computed tomography (PET-CECT) for pulmonary metastasis/synchronous primary lung tumors in head and neck squamous cell cancer (HNSCC) and its association with clinico-radio-pathological factors. Our retrospective study included 837 HNSCC patients from January 2012 to December 2017. Lung nodules were characterized on CT as benign, indeterminate, and metastatic. The true detection rate and statistical significance of associated risk factors were calculated. Risk factors for metastasis were determined using univariate and multivariate logistic regression models. Seventy-five (8.9%) patients had pulmonary metastasis and 3 (0.3%) had second lung primary. Detection rate of pulmonary metastasis by CT was higher (sensitivity-97.3%, specificity-97.2%) as compared to radiograph (sensitivity 49% and specificity 89%). Correlation was found between pulmonary and extra-pulmonary metastasis and N classification (P = 0.01, P = 0.02) and positive low jugular node (P = 0.001, P = 0.001). Using PET-CECT in place of CT costed an extra outlay of 7,033,805 INR (95,551.85 USD) while detecting distant metastasis in only 4 (0.47%) extra cases. Chest CT is a useful pulmonary metastases screening tool in advanced HNSCC patients with reasonable imaging cost as compared to PET-CT.

MR imaging-based risk stratification scoring system to predict clinical outcomes in carotid body tumors.

Objectives: This study aims to evaluate the role of pretherapy MRI in predicting outcomes in carotid body tumors and propose a grading system for high- and low-risk characteristics. Materials and methods: A retrospective observational study of 44 patients with 51 lesions was carried out from year 2005 to 2020. MR images were reviewed for characteristics of carotid body tumor, and a score was given that was correlated with intra- and postoperative findings. The various other classifications and our proposed Mahajan classification were compared with Shamblin's classification. The area under the curve and ROC curves were used to present the accuracy of different predictive models. Results: Our scoring system allotted a score of 0 to 15 on the basis of MRI characteristics, with scores calculated for patients in our study ranging from 0 to 13. Lesions with scores of 0-6 were considered low risk (45%), and scores of 7-15 were regarded as high risk for surgery (55%). The Mahajan classification stages tumors into four grades: I (10%), II (20%), IIIa (8%), and IIIb (62%). The frequency of vascular injury was 50% in category I and 64% in category IIIb. The frequency of cranial nerve injury was 50%, 66%, and 27% in categories I, II, and IIIb. Conclusion: The Mahajan classification of CBTs evaluates high-risk factors like the distance of the tumor from the skull base and the angle of contact with ICA, which form the major predictors of neurovascular damage and morbidity associated with its surgery. Though the Shamblin classification of CBT is the most widely accepted classification, our proposed Mahajan classification system provides an imaging-based alternative to prognosticate surgical candidates preoperatively.

Active Surveillance of Low-Risk Papillary Microcarcinoma of the Thyroid in Indian Scenario: Are we Ready for it? A Narrative Review.

Papillary microcarcinoma (PMC) is defined as papillary thyroid carcinoma (PTC) measuring ≤1 cm, irrespective of the presence or absence of the high-risk features. PMCs without any high-risk features referred to as the low-risk PMCs are generally indolent, and most of them remain latent without progression or with very slow progression. Active surveillance (observation without immediate surgery) could identify the small minority of PMCs that progress and rescue surgery for these PMCs should be effective resulting in no influence on the patients' prognosis than performing immediate surgery which might result in more harm than good due to associated morbidity. So, with proper patient selection, organization, and patient counselling, active surveillance has the potential to be a long-term management strategy for patients with PMC. The recent update of the ATA guidelines (2015) incorporated active surveillance as an option within the management protocol of PTC, making it an considerable rather than an experimental treatment option. The cost for immediate surgery is higher than the medical costs of active surveillance for 10 years in most scenarios. Developing countries like India may have certain limitations like lack of understanding, financial constraints and lack of adequate radiology services; hence, we propose additional recommendations along with standard surveillance strategy.

Prognostic Value of Radiological Extranodal Extension Detected by Computed Tomography for Predicting Outcomes in Patients With Locally Advanced Head and Neck Squamous Cell Cancer Treated With Radical Concurrent Chemoradiotherapy.

Objective: Extra Nodal Extension (ENE) assessment in locally advanced head and neck cancers (LAHNCC) treated with concurrent chemo radiotherapy (CCRT) is challenging and hence the American Joint Committee on Cancer (AJCC) N staging. We hypothesized that radiology-based ENE (rENE) may directly impact outcomes in LAHNSCC treated with radical CCRT. Materials and Methods: Open-label, investigator-initiated, randomized controlled trial (RCT) (2012-2018), which included LAHNSCC planned for CCRT. Patients were randomized 1:1 to radical radiotherapy (66-70 grays) with concurrent weekly cisplatin (30 mg/m2) [cisplatin radiation arm (CRT)] or same schedule of CRT with weekly nimotuzumab (200 mg) [nimotuzumab plus CRT (NCRT)]. A total of 536 patients were accrued and 182 were excluded due to the non-availability of Digital Imaging and Communications in Medicine (DICOM) computed tomography (CT) data. A total of 354 patients were analyzed for rENE. Metastatic nodes were evaluated based on five criteria and further classified as rENE as positive/negative based on three-criteria capsule irregularity with fat stranding, fat invasion, and muscle/vessel invasion. We evaluated the association of rENE and disease-free survival (DFS), loco-regional recurrence-free survival (LRRFS), and overall survival (OS). Results: A total of 244 (68.9%) patients had radiologically metastatic nodes (rN), out of which 140 (57.3%) had rENE. Distribution of rENE was balanced in the two study groups CRT or NCRT (p-value 0.412). The median follow-up period was 39 months (ranging from 35.5 to 42.8 months). Complete response (CR) was seen in 204 (57.6%); incomplete response (IR), i.e., partial response plus stable disease (PR + SD), in 126 (35.6%); and progressive disease (PD) in 24 (6.8%). rENE-positive group had poor survival compared to rENE-negative group 3-year OS (46.7% vs. 63.6%), poor DFS (48.8% vs. 87%), and LRRFS (39.9% vs. 60.4%). rENE positive had 1.71 times increased risk of IR than rENE negative. Overall stage, site, clinical metastatic node (cN), response, and rENE were the significant factors for predicting OS, DFS, and LRRFS on univariate analysis. After making adjustment on multivariate analysis, rENE was an independent prognostic factor for DFS and trending to be significant for OS. Conclusion: Pre-treatment rENE is an independent prognostic marker for survival in patients with LAHNSCC treated radically with CCRT that can be used as a potential predictive marker for response to treatment and hence stratify patients into responders vs. non-responders. We propose the mahajan rENE grading system applicable on CT, magnetic resonance imaging, positron emission tomography-contrast-enhanced CT, and ultrasound.

Imaging for Plasma Cell Dyscrasias: What, When, and How?

Imaging plays a vital role in the diagnosis, response assessment, and follow-up of patients with plasma cell bone disease. The radiologic diagnostic paradigm has thus far evolved with developing technology and availability of better imaging platforms; however, the skewed availability of these imaging modalities in developed vis-à-vis the developing countries along with the lack of uniformity in reporting has led to a consensus on the imaging criteria for diagnosing and response assessment in plasma cell dyscrasia. Therefore, it is imperative for not only the radiologists but also the treating oncologist to be aware of the criteria and appropriate imaging modality to be used in accordance with the clinical question. The review will allow the treating oncologist to answer the following questions on the diagnostic, prognostic, and predictive abilities of various imaging modalities for plasma cell dyscrasia: a) What lesions can look like multiple myeloma (MM) but are not?; b) Does the patient have MM? To diagnose MM in a high-risk SMM patient with clinical suspicion, which modality should be used and why?; c) Is the patient responding to therapy on follow-up imaging once treatment is initiated?; d) To interpret commonly seen complications post-therapy, when is it a disease and when is the expected sequel to treatment? Fractures, red marrow reconversion?; and e) When is the appropriate time to flag a patient for further workup when interpreting MRI spine done for back pain in the elderly? How do we differentiate between commonly seen osteoporosis-related degenerative spine versus marrow infiltrative disorder?

Prognostic Impact of Pattern of Mandibular Involvement in Gingivo-Buccal Complex Squamous Cell Carcinomas: Marrow and Mandibular Canal Staging System.

Purpose: To study the pattern of mandibular involvement and its impact on oncologic outcomes in patients with gingivo-buccal complex squamous cell carcinoma (GBC-SCC) and propose a staging system based on the pattern of bone involvement (MMC: Marrow and mandibular canal staging system) and compare its performance with the 8th edition of the American Joint Committee on Cancer (AJCC8). Methods: This retrospective observational study included treatment-naïve GBC-SCC patients who underwent preoperative computed tomography (CT) imaging between January 1, 2012, and March 31, 2016, at a tertiary care cancer center. Patients with T4b disease with high infratemporal fossa involvement, maxillary erosion, and follow-up of less than a year were excluded. The chi-square or Fisher's exact test was used for descriptive analysis. Kaplan-Meier estimate and log-rank test were performed for survival analysis. Multivariate analysis was done using Cox regression analysis after making adjustments for other prognostic factors. p-Value <0.05 was considered as significant. Based upon the survival analysis with different patterns of bone invasion, a new staging system was proposed "MMC: Marrow and mandibular canal staging system". "Akaike information criterion" (AIC) was used to study the relative fitted model of the various staging (TNM staging-AJCC8) with respect to survival parameters. Results: A total of 1,200 patients were screened; 303 patients were included in the study. On radiology review, mandibular bone was involved in 62% of patients. The pattern of bone involvement was as follows: deep cortical bone erosion (DCBE) in 23%, marrow in 34%, and marrow with the mandibular canal in 43% of patients. Patients with DCBE and no bone involvement (including superficial cortical) had similar survival [disease-free survival (DFS) and locoregional recurrence-free survival (LRRFS)], and this was significantly better than those with marrow with or without mandibular canal involvement (for both DFS and LRRFS). Patients with DCBE were staged using the MMC, and when compared with the AJCC8, the MMC system was better for the prediction of survival outcomes, as AIC values were lower compared with those of the AJCC8. There was a significant association (p = 0.013) between the type of bone involvement and the pattern of recurrence. Conclusions: For GBC-SCC, only marrow with or without mandibular canal involvement is associated with poorer survival outcomes. As compared with the AJCC8, the proposed Mahajan et al. MMC staging system downstages DCBE correlates better with survival outcomes.