Source: https://jubileejoyzirebwa.com/projects/liver-mri-methods/
Canonical: https://jubileejoyzirebwa.com/projects/liver-mri-methods/

[Work](https://jubileejoyzirebwa.com/work/)

# Liver MRI Methods

I built this small introductory project to understand methods developed for patient-level clinical data in the world of multiphasic MRI.

[![Training and validation binary cross-entropy fall across the three U-Net epochs.](https://jubileejoyzirebwa.com/_astro/liver-learning-curve.C8xnkKCY_Z1N3grJ.webp)](https://github.com/JJZirebwa/liver-mri-methods/blob/v0.1.0/docs/assets/learning_curve.png)

Training and validation binary cross-entropy plotted by epoch [Source](https://github.com/JJZirebwa/liver-mri-methods/blob/v0.1.0/docs/assets/learning_curve.png)[Gross et al., LiverHccSeg v1.1](https://doi.org/10.5281/zenodo.8179129)[TCGA-LIHC, The Cancer Imaging Archive](https://doi.org/10.7937/K9/TCIA.2016.IMMQW8UQ)[CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)

## What I was asked

While patient separation matters, image geometry, registration and annotation choices shape the measurement before a model sees it.

## What I did

1. **Geometry**

   Audit phase availability, orientation, spacing and physical grids before comparing voxels across scans.
2. **Registration**

   I compared the dataset-supplied registration with a seeded rigid SimpleITK baseline on three examinations.
3. **Logistic baseline**

   Separate patients before fitting a logistic voxel classifier using four declared dynamic-contrast signal features.
4. **U-Net**

   Run a fixed three-epoch, four-channel 2D U-Net smoke test and review its predictions alongside its numerical scores.
5. **Results**

   Preserve the failed baseline, configuration and result tables so later debugging studies extend the record instead of rewriting it.

## What it changed

The logistic baseline could rank tumour regions strongly while labelling too much liver.

The U-Net produced apparently stronger overlap for one held-out case, but the evaluation units differed and the images told the more important story: broad liver-shaped predictions and false positives in a declared negative examination.

Before making the network larger, the project now needs to test target construction, slice balance and the ways the manual liver mask may have created a shortcut.

Kind of work

Self-directed computational imaging methods project

Dataset

LiverHccSeg multiphasic MRI: 17 examinations with whole-liver annotations

Published release

0.1.0

Technical stack

Python, NumPy, pandas, scikit-learn, SimpleITK and PyTorch

Registration comparison

7 of 9 had higher in-liver mutual-information scores with supplied images

- [Code](https://github.com/JJZirebwa/liver-mri-methods)
- [Baseline study](https://github.com/JJZirebwa/liver-mri-methods/blob/v0.1.0/docs/studies/liverhccseg_baseline_v1.md)

## Further artefacts

[liver-mri-methods docs/studies/liverhccseg\_baseline\_v1.md](https://github.com/JJZirebwa/liver-mri-methods/blob/v0.1.0/docs/studies/liverhccseg_baseline_v1.md)

Method notes, result tables and checksums [Baseline study](https://github.com/JJZirebwa/liver-mri-methods/blob/v0.1.0/docs/studies/liverhccseg_baseline_v1.md)

[Reproducible RNA-seq Service Workflow Next piece of work](https://jubileejoyzirebwa.com/projects/rnaseq-service-workflow/)
