RoboChest:
Interactive Chest X-Ray Teaching and Interpretation Guide

by
Les Folio, DO, MPH, FAOCR, FSCBTMR
Col (ret), USAF
Lead CT Radiologist, Radiology and Imaging Sciences
NIH Clinical Center

Email:
Les.Folio@nih.gov
folio47@gmail.com
Phone:
240 281 8832

Introduction:

The Chest X-Ray (CXR) or chest radiograph remains the most commonly ordered imaging study in medicine yet paradoxically is the most complex. The CXR is the hardest modality for which to learn, recall and master effective and accurate interpretation. The chest radiograph is difficult to interpret, especially in critical care where there are multiple findings. The chest radiograph includes everything in the thorax and provides a high yield, given the low cost and single source. The CXR is also the most commonly utilized imaging study in combat operations.

The following Guide is organized by categories of findings that can be seen on the CXR. This material is presented in a novel format using directory structures to allow students, residents and other providers to dig deeper as needed on a topic at hand. The chunked information was originally designed to link as help pages from a larger project called "ChestWeb," an expert system to help guide image interpretation. This project and a patent (Web-Assisted Diagnosis of Abnormal Densities in Chest Imaging. Provisional filed Jan 2006. Serial# 60/761,777) was abandoned following review of potential impact of availability.

Robochest was updated in September 2017 to supplement an online TB (Tuberculosis) Portal for public use called MXDR-TB DEPOT created for working with cohorts that one can analyze Multiple and Extensively Drug-Resistant TB across four types of data: Clinical, Genomic, CT (Computed Tomography) and chest X-Rays.

https://depot.tbportals.niaid.nih.gov

This work presents a structured lexicon to reproducibly describe radiographic abnormalities detected on plain film CXRs. The lexicon is designed to provide clinically significant differentiation of abnormalities detected. The content is chunked in a directory structure that relates specific combinations of distinct radiographic findings to classes/groupings of pathological etiologies of those findings. Recognizing the individual findings and identifying their combination or lack of combination with other individual findings allows one to create effective differential diagnoses that can then be further evaluated using other imaging procedures and/or non-radiographic clinical information. Included in this work are hundreds of images including x-rays, Computerized Tomography (CT) images, graphics, analogous models and animations to help teach otherwise complex processes and radiographic principles. This material is by no means comprehensive and is designed as a teaching tool. It should not be used for medical diagnosis.

This directory structure method may be helpful to residents in reviewing the general differentials and discussing possibilities with faculty and referring physicians. This method may also be helpful for General Medical Officers (GMOs) in deployed or remote locations without other available references.

Acknowledgements:

This product was created by the Education & Technology (ETI) Support Office at Uniformed Services University. The ETI team methodically organized the lecture materials and other content into sections that would illuminate the complex art of the CXR and then developed the pages and the directory structure. The ETI Support Office is operated by Concurrent Technologies Corporation.

Educational support/ funding:

The ChestWeb project was supported in part by the Henry M. Jackson Foundation (HMJF) and a Corporate Research and Development Award (CRADA) with Expert-24, USU and HMJF. In addition, several intramural grants such as the Dean's Endowment fund helped develop this teaching tool.