vital-ultrasound / preprint2023

:page_facing_up: arxiv preprint 2023
0 stars 0 forks source link

Statistical analysis of participants #12

Closed mxochicale closed 2 years ago

mxochicale commented 2 years ago

Nhat and I discussed the use of potential variables from echo datasets (files ADM.cvs and ADM_DEN_SEP.csv). We have started with a preselection in /proceedings/statistica-analysis. However, we are still unclear on which variables are relevant but this ticket will help to clarify what can be considered to create a statistical analysis of the participants.

The following are references might add better explainability to the collected data:

mxochicale commented 2 years ago

:scroll: Labs R.B., Zolgharni M., Loo J.P. (2021) Echocardiographic Image Quality Assessment Using Deep Neural Networks. In: Papież B.W., Yaqub M., Jiao J., Namburete A.I.L., Noble J.A. (eds) Medical Image Understanding and Analysis. MIUA 2021. Lecture Notes in Computer Science, vol 12722. Springer, Cham. https://doi.org/10.1007/978-3-030-80432-9_36

The study population consisted of a random sample of (PACS2-Dataset) 1,039 Echocar- diographic studies from patients with age ranges from 17 and 85 years, who were recruited from patients who had undergone echocardiography with Imperial College Healthcare NHS Trust. The acquisition of the images had been completed by experi- enced Echocardiographers using ultrasound equipment from GE healthcare (Vivid.1) and Philips Healthcare (iE33 xMatrix) manufacturers according to the standard protocols.

https://github.com/vital-ultrasound/2022-echocardiography-proceedings/tree/10-new-literature/proceedings/references/2021-labs-miua-SPRINGER/article.pdf

:scroll: Zhang, J., Gajjala, S., Agrawal, P., Tison, G. H., Hallock, L. A., Beussink-Nelson, L., ... & Deo, R. C. (2018). Fully automated echocardiogram interpretation in clinical practice: feasibility and diagnostic accuracy. Circulation, 138(16), 1623-1635.**

A total of 14 035 echocardiograms were used for this proj- ect. Echocardiograms each consist of a collection of video and still images collected on a single patient at a single time. These studies span a period of 10 years and were acquired using diverse echocardiography devices (3 manufacturers and 10 models; https://doi.org/10.1161/CIRCULATIONAHA.118.034338

Screenshot from 2022-02-15 08-32-05

:scroll: Melissa C. Brindise, Brett A. Meyers, Shelby Kutty, Pavlos P. Vlachos "Unsupervised Segmentation of B-Mode Echocardiograms" 2020 in arvix

The cohort consisted of 66 subjects: 4 dilated cardiomyopathy (DCM), 20 hypertrophic cardiomyopathy (HCM), and 42 age- matched controls (Normal). Cohort demographics are provided in Table 1 and function indices are provided in Table 2. https://arxiv.org/abs/2010.11816

Screenshot from 2022-02-17 17-00-19

:scroll: Ghorbani, A., Ouyang, D., Abid, A. et al. Deep learning interpretation of echocardiograms. npj Digit. Med. 3, 10 (2020). https://doi.org/10.1038/s41746-019-0216-8

Table 1. Baseline characteristics of patients in the training and test datasets Screenshot from 2022-03-28 13-11-53

mxochicale commented 2 years ago

In our weekly meeting of 15 Feb 2022, we discussed our little understanding of patient health conditions when echos were acquired to which the following table would help to distil what we are hypothesising about patient state of health and the challenges of performing echocardiography in the ICU. Q: Do we know the following variables from our datasets?

Patient Gender Age Height Weight Disease Cardiac Disease Patient conditions in the ICU (cannot move, with ventilator, others) ...
P00 M/F
... ...
PNN M/F

Potential Cardiac Diseases:

nhatpth commented 2 years ago

Hi Miguel, yes, we have collect the information that you mentioned (Gender | Age | Height | Weight | Disease | Cardiac Disease | Patient conditions in the ICU (cannot move, with ventilator, others). In our patients, only few of them having cardiac diseases.

mxochicale commented 2 years ago

Addressed here patient_demographics_and_diseases.ipynb