Foundational Segmentation Models and Clinical Data Mining Enable Accurate Computer Vision for Lung Cancer.

This study aims to assess the effectiveness of integrating Segment Anything Model (SAM) and its variant MedSAM into the automated mining, object detection, and segmentation (MODS) methodology for developing robust lung cancer detection and segmentation models without post hoc labeling of training images. In a retrospective analysis, 10,000 chest computed tomography scans from patients with lung cancer were mined. Line measurement annotations were converted to bounding boxes, excluding boxes < 1 cm or > 7 cm. The You Only Look Once object detection architecture was used for teacher-student learning to label unannotated lesions on the training images. Subsequently, a final tumor detection model was trained and employed with SAM and MedSAM for tumor segmentation. Model performance was assessed on a manually annotated test dataset, with additional evaluations conducted on an external lung cancer dataset before and after detection model fine-tuning. Bootstrap resampling was used to calculate 95% confidence intervals. Data mining yielded 10,789 line annotations, resulting in 5403 training boxes. The baseline detection model achieved an internal F1 score of 0.847, improving to 0.860 after self-labeling. Tumor segmentation using the final detection model attained internal Dice similarity coefficients (DSCs) of 0.842 (SAM) and 0.822 (MedSAM). After fine-tuning, external validation showed an F1 of 0.832 and DSCs of 0.802 (SAM) and 0.804 (MedSAM). Integrating foundational segmentation models into the MODS framework results in high-performing lung cancer detection and segmentation models using only mined clinical data. Both SAM and MedSAM hold promise as foundational segmentation models for radiology images.

Body mass index and adiposity influence responses to immune checkpoint inhibition in endometrial cancer.

Background: Obesity is the foremost risk factor in the development of endometrial cancer (EC). However, the impact of obesity on the response to immune checkpoint inhibitors (ICI) in EC remains poorly understood. This retrospective study investigates the association between body mass index (BMI), body fat distribution, and clinical and molecular characteristics of EC patients treated with ICI. Methods: We analyzed progression-free survival (PFS) and overall survival (OS) in EC patients treated with ICI, categorized by BMI, fat mass distribution, and molecular subtypes. Incidence of immune-related adverse events (irAE) after ICI was also assessed based on BMI status. Results: 524 EC patients were included in the study. Overweight and obese patients exhibited a significantly prolonged PFS and OS compared to normal BMI patients after treatment with ICI. Multivariable Cox regression analysis confirmed the independent association of overweight and obesity with improved PFS and OS. Elevated visceral adipose tissue (VAT) was identified as a strong independent predictor for improved PFS to ICI. Associations between obesity and OS/PFS were particularly significant in the copy number-high/TP53abnormal (CN-H/TP53abn) EC molecular subtype. Finally, obese patients demonstrated a higher irAE rate compared to normal BMI individuals. Conclusion: Obesity is associated with improved outcomes to ICI in EC patients and a higher rate of irAEs. This association is more pronounced in the CN-H/TP53abn EC molecular subtype. Funding: NIH/NCI Cancer Center Support Grant P30CA008748 (MSK). K08CA266740 and MSK Gerstner Physician Scholars Program (J.C.O). RUCCTS Grant #UL1 TR001866 (N.G-B and C.S.J). Cycle for survival and Breast Cancer Research Foundation grants (B.W).

Body mass index and adiposity influence responses to immune checkpoint inhibition in endometrial cancer.

BACKGROUNDObesity is the foremost risk factor in the development of endometrial cancer (EC). However, the impact of obesity on the response to immune checkpoint inhibitors (ICI) in EC remains poorly understood. This retrospective study investigates the association among BMI, body fat distribution, and clinical and molecular characteristics of EC patients treated with ICI.METHODSWe analyzed progression-free survival (PFS) and overall survival (OS) in EC patients treated with ICI, categorized by BMI, fat-mass distribution, and molecular subtypes. Incidence of immune-related adverse events (irAEs) after ICI was also assessed based on BMI status.RESULTS524 EC patients were included in the study. Overweight and obese patients exhibited a significantly prolonged PFS and OS compared with normal BMI patients after treatment with ICI. Multivariable Cox's regression analysis confirmed the independent association of overweight and obesity with improved PFS and OS. Elevated visceral adipose tissue (VAT) was identified as a strong independent predictor for improved PFS to ICI. Associations between obesity and OS/PFS were particularly significant in the copy number-high/TP53abnormal (CN-H/TP53abn) EC molecular subtype. Finally, obese patients demonstrated a higher irAE rate compared with normal BMI individuals.CONCLUSIONObesity is associated with improved outcomes to ICI in EC patients and a higher rate of irAEs. This association is more pronounced in the CN-H/TP53abn EC molecular subtype.FUNDINGNIH/NCI Cancer Center; MSK Gerstner Physician Scholars Program; National Center for Advancing Translational Sciences (NCATS); Cycle for Survival; Breast Cancer Research Foundation.

Biomarkers associated with pulmonary exacerbations in a randomized trial of nintedanib for radiation pneumonitis.

BACKGROUND AND PURPOSE: Radiation pneumonitis (RP) is a common side effect of thoracic radiotherapy and often has a long course characterized by acute exacerbations and progression to permanent lung fibrosis. There are no validated biomarkers of prognosis in patients diagnosed with RP. MATERIALS AND METHODS: We analyzed a time course of serum chemokines, cytokines, and other proteins from patients with grade 2+ RP in a randomized clinical trial of a steroid taper plus nintedanib, a multiple tyrosine kinase inhibitor, versus placebo plus a steroid taper for the treatment of RP. Weighted gene correlation network analysis (WGCNA) and univariable zero inflated Poisson models were used to identify groups of correlated analytes and their associations with clinical outcomes. RESULTS: Thirty enrolled patients had biomarker data available, and 17 patients had enough analytes tested for network analysis. WGNCA identified ten analytes, including transforming growth factor beta-1 (TGF-ß1), monocyte chemoattractant protein-1 (MCP-1), and platelet-derived growth factor (PDGF), that in aggregate were correlated with the occurrence of pulmonary exacerbations (p = 0.008), the total number of acute pulmonary exacerbations (p = 0.002), and treatment arm (p = 0.036). By univariable analysis, an increase in rate of change of two components of the RP module were associated with an increased incidence rate of pulmonary exacerbations: interleukin 5 (IL-5, incidence rate ratio (IRR) 1.02, 95% CI 1.01-1.04, p = 0.002), and tumor necrosis factor superfamily 12 (TNFSF12, IRR 1.06, CI 1-1.11, p = 0.036). An increased slope of epidermal growth factor (EGF) was associated with a decreased incidence rate of exacerbations (IRR 0.94, CI 0.89-1, p = 0.036). CONCLUSION: We identified a panel of serum biomarkers that showed association with nintedanib treatment and acute pulmonary exacerbations in patients with RP. A confirmatory study will be needed to validate this panel for use as a prognostic tool in patients with RP.

Postoperative CT surveillance in the evaluation of local recurrence after sub-lobar resection of neoplastic lesions of the lung.

OBJECTIVE: As indications for sub-lobar resections increase, it will become more important to identify risk factors for postsurgical recurrence. We investigated retrospectively the association between local recurrence after sub-lobar resection of neoplastic lung lesions and pre- and post-operative CT imaging and pathologic features. MATERIALS AND METHODS: We reviewed retrospectively neoplastic lung lesions with postoperative chest CT surveillance of sub-lobar resections in 2006-2016. We defined "suspicious" findings as nodularity ≥3 mm or soft tissue thickening ≥4 mm along the suture line and/or progression and explored their association with local recurrence. Primary lung cancer stage, tumoral invasion of lymphatics, visceral pleura or large vessels, bronchial and vascular margin distance were also assessed. RESULTS: Our study group included 45 cases of sub-lobar resection took for either primary (n = 37) or metastatic (n = 8) lung tumors. Local recurrence was observed in 16 of those patients. New nodularity ≥3 mm or soft tissue thickening ≥4 mm along the suture line on surveillance CT was significantly associated with local recurrence (p = 0.037). Additionally, solid nodule (p = 0.005), age at surgery ≤60 years (p = 0.006), two or more sites of invasion (p < 0.0001) and poor histologic differentiation (p = 0.0001) were also significantly associated with local tumor recurrence. Of 16 patients with surveillance post-surgical PET-CT, 15 had elevated FDG uptake. CONCLUSION: The postoperative changes along the suture line should follow a predictable time course demonstrating a pattern of stability, thinning or resolution on CT surveillance. New or increasing postoperative nodularity ≥3 mm or soft tissue thickening ≥4 mm along the suture line requires close diagnostic work-up. Surgical pathology characteristics added prognostic value on postoperative recurrence surveillance.

Conventional and radiomic features to predict pathology in the preoperative assessment of anterior mediastinal masses.

OBJECTIVES: The aim of this study was to differentiate benign from malignant tumors in the anterior mediastinum based on computed tomography (CT) imaging characteristics, which could be useful in preoperative planning. Additionally, our secondary aim was to differentiate thymoma from thymic carcinoma, which could guide the use of neoadjuvant therapy. MATERIALS AND METHODS: Patients referred for thymectomy were retrospectively selected from our database. Twenty-five conventional characteristics were evaluated by visual analysis, and 101 radiomic features were extracted from each CT. In the step of model training, we applied support vector machines to train classification models. Model performance was assessed using the area under the receiver operating curves (AUC). RESULTS: Our final study sample comprised 239 patients, 59 (24.7 %) with benign mediastinal lesions and 180 (75.3 %) with malignant thymic tumors. Among the malignant masses, there were 140 (58.6 %) thymomas, 23 (9.6 %) thymic carcinomas, and 17 (7.1 %) non-thymic lesions. For the benign versus malignant differentiation, the model that integrated both conventional and radiomic features achieved the highest diagnostic performance (AUC = 0.715), in comparison to the conventional (AUC = 0.605) and radiomic-only (AUC = 0.678) models. Similarly, regarding thymoma versus thymic carcinoma differentiation, the model that integrated both conventional and radiomic features also achieved the highest diagnostic performance (AUC = 0.810), in comparison to the conventional (AUC = 0.558) and radiomic-only (AUC = 0.774) models. CONCLUSION: CT-based conventional and radiomic features with machine learning analysis could be useful for predicting pathologic diagnoses of anterior mediastinal masses. The diagnostic performance was moderate for differentiating benign from malignant lesions and good for differentiating thymomas from thymic carcinomas. The best diagnostic performance was achieved when both conventional and radiomic features were integrated in the machine learning algorithms.

Multimodal integration of radiology, pathology and genomics for prediction of response to PD-(L)1 blockade in patients with non-small cell lung cancer.

Immunotherapy is used to treat almost all patients with advanced non-small cell lung cancer (NSCLC); however, identifying robust predictive biomarkers remains challenging. Here we show the predictive capacity of integrating medical imaging, histopathologic and genomic features to predict immunotherapy response using a cohort of 247 patients with advanced NSCLC with multimodal baseline data obtained during diagnostic clinical workup, including computed tomography scan images, digitized programmed death ligand-1 immunohistochemistry slides and known outcomes to immunotherapy. Using domain expert annotations, we developed a computational workflow to extract patient-level features and used a machine-learning approach to integrate multimodal features into a risk prediction model. Our multimodal model (area under the curve (AUC) = 0.80, 95% confidence interval (CI) 0.74-0.86) outperformed unimodal measures, including tumor mutational burden (AUC = 0.61, 95% CI 0.52-0.70) and programmed death ligand-1 immunohistochemistry score (AUC = 0.73, 95% CI 0.65-0.81). Our study therefore provides a quantitative rationale for using multimodal features to improve prediction of immunotherapy response in patients with NSCLC using expert-guided machine learning.

Barotrauma in COVID 19: Incidence, pathophysiology, and effect on prognosis.

OBJECTIVES: To investigate the incidence, risk factors, and outcomes of barotrauma (pneumomediastinum and subcutaneous emphysema) in mechanically ventilated COVID-19 patients. To describe the chest radiography patterns of barotrauma and understand the development in relation to mechanical ventilation and patient mortality. METHODS: We performed a retrospective study of 363 patients with COVID-19 from March 1 to April 8, 2020. Primary outcomes were pneumomediastinum or subcutaneous emphysema with or without pneumothorax, pneumoperitoneum, or pneumoretroperitoneum. The secondary outcomes were length of intubation and death. In patients with pneumomediastinum and/or subcutaneous emphysema, we conducted an imaging review to determine the timeline of barotrauma development. RESULTS: Forty three out of 363 (12%) patients developed barotrauma radiographically. The median time to development of either pneumomediastinum or subcutaneous emphysema was 2 days (IQR 1.0-4.5) after intubation and the median time to pneumothorax was 7 days (IQR 2.0-10.0). The overall incidence of pneumothorax was 28/363 (8%) with an incidence of 17/43 (40%) in the barotrauma cohort and 11/320 (3%) in those without barotrauma (p ≤ 0.001). In total, 257/363 (71%) patients died with an increase in mortality in those with barotrauma 33/43 (77%) vs. 224/320 (70%). When adjusting for covariates, barotrauma was associated with increased odds of death (OR 2.99, 95% CI 1.25-7.17). CONCLUSION: Barotrauma is a frequent complication of mechanically ventilated COVID-19 patients. In comparison to intubated COVID-19 patients without barotrauma, there is a higher rate of pneumothorax and an increased risk of death.

Evaluation of interstitial lung disease: An algorithmic review using ILD-RADS.

Interstitial lung diseases (ILDs) may present a diagnostic dilemma due to their many classifications and overlapping imaging findings. In this review, we present an algorithmic approach for assessing ILDs based on identifying and understanding key imaging features to aid in narrowing a differential diagnosis or reaching a specific diagnosis. We use the recently introduced Interstitial Lung Disease Reporting And Data System (ILD-RADS) as a framework for our discussion.

Intra- and inter-reader agreement of iRECIST and RECIST 1.1 criteria for the assessment of tumor response in patients receiving checkpoint inhibitor immunotherapy for lung cancer.

OBJECTIVES: To investigate the inter- and intra-reader agreement of immune Response Evaluation Criteria in Solid Tumors (iRECIST) and Response Evaluation Criteria in Solid Tumors version 1.1 (RECIST 1.1) in patients with lung cancer treated with immunotherapy. MATERIALS AND METHODS: This retrospective study included 85 patients with lung cancer treated with PD-1 blockade. Four radiologists evaluated computed topography (CT) scans before and after initiation of immunotherapy using iRECIST and RECIST 1.1. Weighted kappa (k) with equal weights was used to assess the intra-reader agreement between 2 repeated reads on overall response at all time points, best overall response, and the response at the time point of progression, as well as the intra-reader agreement between iRECIST and RECIST. The inter-reader agreement was calculated using Light's kappa. RESULTS: Intra-reader agreement for overall response at all time points, best overall response, and time point of progression was substantial to almost perfect for both iRECIST and RECIST 1.1 (k = 0.651-0.983). Inter-reader agreement was substantial for iRECIST (κ = 0.657-0.742) while RECIST 1.1 was moderate to substantial (κ = 0.587-0.686). The level of inter-reader agreement was not higher on repeat read for iRECIST (κ = 0.677-0.709 and κ = 0.657-0.742 for first and second read, respectively) as well as for RECIST 1.1 (κ = 0.587-0.659 and κ = 0.633-0.686 for first and second read, respectively). Almost perfect agreement was observed between RECIST 1.1 and iRECIST at first (κ = 0.813-0.923) and second read (κ = 0.841-0.912). CONCLUSION: The inter- and intra-reader agreement of iRECIST is high and similar to RECIST 1.1 in patients with lung cancer treated with immunotherapy.

Loss of retinoid X receptor gamma subunit impairs group 1 mGluR mediated electrophysiological responses and group 1 mGluR dependent behaviors.

Retinoid X receptors are members of the nuclear receptor family that regulate gene expression in response to retinoic acid and related ligands. Group 1 metabotropic glutamate receptors are G-protein coupled transmembrane receptors that activate intracellular signaling cascades in response to the neurotransmitter, glutamate. These two classes of molecules have been studied independently and found to play important roles in regulating neuronal physiology with potential clinical implications for disorders such as depression, schizophrenia, Parkinson's and Alzheimer's disease. Here we show that mice lacking the retinoid X receptor subunit, RXRγ, exhibit impairments in group 1 mGluR-mediated electrophysiological responses at hippocampal Schaffer collateral-CA1 pyramidal cell synapses, including impaired group 1 mGluR-dependent long-term synaptic depression (LTD), reduced group 1 mGluR-induced calcium release, and loss of group 1 mGluR-activated voltage-sensitive currents. These animals also exhibit impairments in a subset of group 1 mGluR-dependent behaviors, including motor performance, spatial object recognition, and prepulse inhibition. Together, these observations demonstrate convergence between the RXRγ and group 1 mGluR signaling pathways that may function to coordinate their regulation of neuronal activity. They also identify RXRγ as a potential target for the treatment of disorders in which group 1 mGluR signaling has been implicated.

Portable Chest Radiography as an Exclusionary Test for Adverse Clinical Outcomes During the COVID-19 Pandemic.

BACKGROUND: Chest radiography (CXR) often is performed in the acute setting to help understand the extent of respiratory disease in patients with COVID-19, but a clearly defined role for negative chest radiograph results in assessing patients has not been described. RESEARCH QUESTION: Is portable CXR an effective exclusionary test for future adverse clinical outcomes in patients suspected of having COVID-19? STUDY DESIGN AND METHODS: Charts of consecutive patients suspected of having COVID-19 at five EDs in New York City between March 19, 2020, and April 23, 2020, were reviewed. Patients were categorized based on absence of findings on initial CXR. The primary outcomes were hospital admission, mechanical ventilation, ARDS, and mortality. RESULTS: Three thousand two hundred forty-five adult patients, 474 (14.6%) with negative initial CXR results, were reviewed. Among all patients, negative initial CXR results were associated with a low probability of future adverse clinical outcomes, with negative likelihood ratios of 0.27 (95% CI, 0.23-0.31) for hospital admission, 0.24 (95% CI, 0.16-0.37) for mechanical ventilation, 0.19 (95% CI, 0.09-0.40) for ARDS, and 0.38 (95% CI, 0.29-0.51) for mortality. Among the subset of 955 patients younger than 65 years and with a duration of symptoms of at least 5 days, no patients with negative CXR results died, and the negative likelihood ratios were 0.17 (95% CI, 0.12-0.25) for hospital admission, 0.09 (95% CI, 0.02-0.36) for mechanical ventilation, and 0.09 (95% CI, 0.01-0.64) for ARDS. INTERPRETATION: Initial CXR in adult patients suspected of having COVID-19 is a strong exclusionary test for hospital admission, mechanical ventilation, ARDS, and mortality. The value of CXR as an exclusionary test for adverse clinical outcomes is highest among young adults, patients with few comorbidities, and those with a prolonged duration of symptoms.

In vitro angiographic comparison of the flow-diversion performance of five neurovascular stents.

Background and purpose Data differentiating flow diversion properties of commercially available low- and high-porosity stents are limited. This in vitro study applies angiographic analysis of intra-aneurysmal flow to compare the flow-diversion performance of five neurovascular devices in idealized sidewall and bifurcation aneurysm models. Methods Five commercial devices (Enterprise, Neuroform, LVIS, FRED, and Pipeline) were implanted in silicone sidewall and bifurcation aneurysm models under physiological average flow of blood analog fluid. High-speed angiographic images were acquired pre- and post-device implantation and contrast concentration-time curves within the aneurysm were recorded. The curves were quantified with five parameters to assess changes in contrast transport, and thus aneurysm hemodynamics, due to each device. Results Inter-device flow-diversion performance was more easily distinguished in the sidewall model than the bifurcation model. There were no obvious overall statistical trends in the bifurcation parameters but the Pipeline performed marginally better than the other devices. In the sidewall geometry, overall evidence suggests that the LVIS performed better than the Neuroform and Enterprise. The Pipeline and FRED devices were statistically superior to the three stents and Pipeline was superior to FRED in all sidewall parameters evaluated. Conclusions Based on this specific set of experiments, lower-porosity flow diverters perform significantly better in reducing intra-aneurysmal flow activity than higher-porosity stents in sidewall-type geometries. The LVIS device is potentially a better flow diverter than the Neuroform and Enterprise devices, while the Pipeline is potentially better than the FRED.

Neuralized1 activates CPEB3: a function for nonproteolytic ubiquitin in synaptic plasticity and memory storage.

The cytoplasmic polyadenylation element-binding protein 3 (CPEB3), a regulator of local protein synthesis, is the mouse homolog of ApCPEB, a functional prion protein in Aplysia. Here, we provide evidence that CPEB3 is activated by Neuralized1, an E3 ubiquitin ligase. In hippocampal cultures, CPEB3 activated by Neuralized1-mediated ubiquitination leads both to the growth of new dendritic spines and to an increase of the GluA1 and GluA2 subunits of AMPA receptors, two CPEB3 targets essential for synaptic plasticity. Conditional overexpression of Neuralized1 similarly increases GluA1 and GluA2 and the number of spines and functional synapses in the hippocampus and is reflected in enhanced hippocampal-dependent memory and synaptic plasticity. By contrast, inhibition of Neuralized1 reduces GluA1 and GluA2 levels and impairs hippocampal-dependent memory and synaptic plasticity. These results suggest a model whereby Neuralized1-dependent ubiquitination facilitates hippocampal plasticity and hippocampal-dependent memory storage by modulating the activity of CPEB3 and CPEB3-dependent protein synthesis and synapse formation.