Public repository for the unpublished full manuscript and legacy analysis code for Pulmonary Artery Size on CT and Mortality Risk: A Dose-Response Analysis, presented as a CHEST 2023 abstract.
The work evaluates whether pulmonary artery diameter (PAd) and pulmonary artery to ascending aorta ratio (PA:AA) have dose-response associations with mortality after emergency-department computed tomography pulmonary angiography (CTPA), and whether age/sex normalization changes prognostic interpretation.
This repository is associated with a conference abstract and an unpublished author manuscript draft. There is no final peer-reviewed journal article for this dose-response analysis.
- CHEST 2023 abstract: Pulmonary Artery Size on CT and Mortality Risk: A Dose-Response Analysis, CHEST 2023 annual meeting abstracts, volume 164, issue 4 supplement, pages A6004-A6005.
- Public CV/abstract page: Brian W. Locke abstracts and presentations.
- Unpublished author draft: manuscript/ctpa-dose-response-unpublished-author-draft.md.
- Source DOCX for provenance: manuscript/source/2023-10-10-dose-response-letter.docx.
- Legacy rendered notebook exports: legacy-rendered-artifacts-2026-06-04.
Related background paper, not the primary citation target for this repository:
- Scarpato BM, Locke BW, Bledsoe J, et al. The association between pulmonary artery enlargement and mortality in an Emergency Department population undergoing computed tomography pulmonary angiography. Pulmonary Circulation. 2023;13(2):e12225. DOI: 10.1002/pul2.12225.
Authors listed on the CHEST abstract and manuscript draft: Brian W. Locke, Brittany M. Scarpato, Joseph Bledsoe, Daniel B. Knox, Karen Conner, Gregory J. Stoddard, Meghan M. Cirulis, C. Gregory Elliott, and Mark W. Dodson.
Maintainer: Brian W. Locke, ORCID 0000-0002-3588-5238.
The draft states that Brian W. Locke was supported by NIH Ruth L. Kirschstein National Research Service Award 5T32HL105321. It also notes unrelated funding through an American Thoracic Society program supported by ResMed, Philips Respironics, and Fisher & Paykel Healthcare.
The analytic datasets are restricted clinical imaging/EHR-derived files and are not public. Do not commit patient-level data, row-level derived datasets, identifiers, local registry files, or collaborator-only workbooks.
Expected local-only inputs:
data/private/final_noempi.dtadata/private/cleaned_noempi.dtadata/private/ACCT EMPI BMI Weight Height.csv
The noempi names indicate prior removal of direct EMPI identifiers, but these files still represent patient-level clinical data and must remain restricted.
| Path | Purpose |
|---|---|
Data_preprocess.do |
Cleans restricted source data and merges supplemental age/BMI/body-size fields. |
Age-Metric Analysis.do |
Main age/sex normalization and survival-analysis workflow for the dose-response manuscript. |
PA Body-size Analysis.do |
Exploratory analysis of PA, AA, body size, ratios, and mortality prediction. |
BL CombofilewithoutEMPI.do |
Older combined preprocessing/analysis script retained for legacy provenance. |
Unit Tests.do, calc_ibs.py, define_calc_ibs.do |
Experimental integrated Brier score test/prototype materials. |
age-sex-norm-letter.ipynb, ratio_val_analysis.ipynb |
Historical exploratory notebooks with cleared outputs. |
manuscript/ |
Unpublished author manuscript draft and source DOCX. |
data_dictionary.md, data_dictionary.csv |
Machine-readable and human-readable variable documentation. |
llms.txt, AGENTS.md, CITATION.cff |
Machine-readable repository orientation and citation metadata. |
Third-party reference PDFs and generated HTML/PDF notebook exports are intentionally not tracked on the branch tip. Use DOI links and the legacy rendered-artifacts release instead.
Run Stata from the repository root. Full execution requires local restricted inputs.
stata-mp -b do "Data_preprocess.do" "data/private" "outputs/stata"
stata-mp -b do "Age-Metric Analysis.do" "data/private" "outputs/stata"
stata-mp -b do "PA Body-size Analysis.do" "data/private" "outputs/stata"Default arguments are data/private for inputs and outputs/stata for generated outputs. The older combined workflow can be run similarly:
stata-mp -b do "BL CombofilewithoutEMPI.do" "data/private" "outputs/stata"Generated logs, copied do-files, tables, graphs, and derived .dta files are written under outputs/stata/<date>/ and ignored by git.
| Component | Requirement |
|---|---|
| Stata | Tested/developed with Stata 17-18 syntax. |
| Stata packages | missings, mdesc, table1_mc, coefplot, outreg2, mkspline2, xblc, stcoxkm, cleanplots, schemepack; some exploratory commands may also need palettes, white_tableau, or related schemes. |
| Python notebooks | Install requirements.txt for notebook support and the integrated Brier score prototype. |
| Restricted data | Not included; supply compatible local files under data/private/. |
Expected generated outputs include:
- cleaned or merged local
.dtafiles; - Stata logs and copied do-files;
- table exports such as age-stratified baseline characteristics;
- survival curves, spline plots, quantile-regression plots, and PA/AA/PA:AA figures.
Historical rendered notebook exports are archived in the GitHub release named legacy-rendered-artifacts-2026-06-04. They are retained for provenance and reading convenience, not as canonical reproducible outputs.
If citing the scholarly work, cite the CHEST abstract:
Locke BW, Scarpato B, Bledsoe J, Knox D, Conner K, Stoddard GJ, Cirulis MM, Elliott CG, Dodson MW. Pulmonary Artery Size on CT and Mortality Risk: A Dose-Response Analysis. CHEST. 2023;164(4 Suppl):A6004-A6005. doi:10.1016/j.chest.2023.07.3871.
If citing repository code or draft materials, cite the repository URL, commit SHA, and CITATION.cff.
Repository code and original documentation are released under the MIT License. Restricted clinical data, third-party reference papers, publisher-formatted PDFs, and external datasets are excluded. The unpublished author draft is included for scholarly indexing and review; verify authorship, permissions, and journal/preprint policy before redistributing outside the repository context.
Use GitHub issues or pull requests for public repository questions. Do not post PHI, patient-level data, local registry files, or collaborator-only materials in issues or pull requests.