Developing radiology diagnostic tools for pulmonary fibrosis using machine learning methods.

BACKGROUND: Accurate and prompt diagnosis of the different patterns for pulmonary fibrosis is essential for patient management. However, accurate diagnosis of the specific pattern is challenging due to overlapping radiographic characteristics. MATERIALS AND METHODS: We conducted a retrospective chart review utilizing two machine learning methods, classification and regression tree and Bayesian additive regression tree, to select the most important radiographic features for diagnosing the three most common fibrosis patterns and created an online diagnostic app for convenient implementation. RESULTS: Four hundred patients (median age of 67 with inter quartile range 58-73; 200 males) were included in the study. Peripheral distribution, homogeneity, lower lobe predominance and mosaic attenuation of fibrosis are the four most important features identified. Bayesian additive regression tree demonstrates better performance than classification and regression tree in diagnosis prediction and provides the predicted probability of each diagnosis with uncertainty intervals for each combination of features. CONCLUSION: The model and app built with Bayesian additive regression tree can be used as an effective tool in assisting radiologists in the diagnostic process of pulmonary fibrosis pattern recognition.

Imaging the Side Effects of CAR T Cell Therapy: A Primer for the Practicing Radiologist.

Chimeric antigen receptor (CAR) T cell therapy is a revolutionary form of immunotherapy that has proven to be efficacious in the treatment of many hematologic cancers. CARs are modified T lymphocytes that express an artificial receptor specific to a tumor-associated antigen. These engineered cells are then reintroduced to upregulate the host immune responses and eradicate malignant cells. While the use of CAR T cell therapy is rapidly expanding, little is known about how common side effects such as cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity (ICANS) present radiographically. Here we provide a comprehensive review of how side effects present in different organ systems and how they can be optimally imaged. Early and accurate recognition of the radiographic presentation of these side effects is critical to the practicing radiologist and their patients so that these side effects can be promptly identified and treated.

The Macklin effect closely correlates with pneumomediastinum in acutely ill intubated patients with COVID-19 infection.

PURPOSE: Patients with COVID-19 infection are frequently found to have pulmonary barotrauma. Recent work has identified the Macklin effect as a radiographic sign that often occurs in patients with COVID-19 and may correlate with barotrauma. METHODS: We evaluated chest CT scans in COVID-19 positive mechanically ventilated patients for the Macklin effect and any type of pulmonary barotrauma. Patient charts were reviewed to identify demographic and clinical characteristics. RESULTS: The Macklin effect on chest CT scan was identified in a total of 10/75 (13.3%) COVID-19 positive mechanically ventilated patients; 9 developed barotrauma. Patients with the Macklin effect on chest CT scan had a 90% rate of pneumomediastinum (p < 0.001) and a trend toward a higher rate of pneumothorax (60%, p = 0.09). Pneumothorax was most frequently omolateral to the site of the Macklin effect (83.3%). CONCLUSION: The Macklin effect may be a strong radiographic biomarker for pulmonary barotrauma, most strongly correlating with pneumomediastinum. Studies in ARDS patients without COVID-19 are needed to validate this sign in a broader population. If validated in a broad population, future critical care treatment algorithms may include the Macklin sign for clinical decision making and prognostication.