Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

PARASIDE: An Automatic Paranasal Sinus Segmentation and Structure Analysis Tool for MRI

Hendrik Möller, Lukas Krautschick, Matan Atad, Robert Graf, Chia-Jung Busch, Achim Beule, Christian Scharf, Lars Kaderali, Bjoern Menze, Daniel Rueckert, Jan Kirschke, Fabian Schwitzing

TL;DR

PARASIDE, an automatic tool for segmenting air and soft tissue volumes of the structures of the sinus maxillaris, frontalis, sphenodalis and ethmoidalis in T1 MRI, is introduced, which is the first automated whole nasal segmentation of 16 structures, and capable of calculating medical relevant features such as the Lund-Mackay score.

Abstract

Chronic rhinosinusitis (CRS) is a common and persistent sinus imflammation that affects 5 - 12\% of the general population. It significantly impacts quality of life and is often difficult to assess due to its subjective nature in clinical evaluation. We introduce PARASIDE, an automatic tool for segmenting air and soft tissue volumes of the structures of the sinus maxillaris, frontalis, sphenodalis and ethmoidalis in T1 MRI. By utilizing that segmentation, we can quantify feature relations that have been observed only manually and subjectively before. We performed an exemplary study and showed both volume and intensity relations between structures and radiology reports. While the soft tissue segmentation is good, the automated annotations of the air volumes are excellent. The average intensity over air structures are consistently below those of the soft tissues, close to perfect separability. Healthy subjects exhibit lower soft tissue volumes and lower intensities. Our developed system is the first automated whole nasal segmentation of 16 structures, and capable of calculating medical relevant features such as the Lund-Mackay score.

PARASIDE: An Automatic Paranasal Sinus Segmentation and Structure Analysis Tool for MRI

TL;DR

PARASIDE, an automatic tool for segmenting air and soft tissue volumes of the structures of the sinus maxillaris, frontalis, sphenodalis and ethmoidalis in T1 MRI, is introduced, which is the first automated whole nasal segmentation of 16 structures, and capable of calculating medical relevant features such as the Lund-Mackay score.

Abstract

Chronic rhinosinusitis (CRS) is a common and persistent sinus imflammation that affects 5 - 12\% of the general population. It significantly impacts quality of life and is often difficult to assess due to its subjective nature in clinical evaluation. We introduce PARASIDE, an automatic tool for segmenting air and soft tissue volumes of the structures of the sinus maxillaris, frontalis, sphenodalis and ethmoidalis in T1 MRI. By utilizing that segmentation, we can quantify feature relations that have been observed only manually and subjectively before. We performed an exemplary study and showed both volume and intensity relations between structures and radiology reports. While the soft tissue segmentation is good, the automated annotations of the air volumes are excellent. The average intensity over air structures are consistently below those of the soft tissues, close to perfect separability. Healthy subjects exhibit lower soft tissue volumes and lower intensities. Our developed system is the first automated whole nasal segmentation of 16 structures, and capable of calculating medical relevant features such as the Lund-Mackay score.
Paper Structure (13 sections, 6 figures, 3 tables)

This paper contains 13 sections, 6 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Materials and methods
  3. Data
  4. Segmentation Approach
  5. Downstream Utility
  6. Experiments
  7. Results and Discussion
  8. Segmentation Performance
  9. Feature Analysis
  10. Discussion
  11. Conclusion
  12. Tables with captions
  13. Appendix

Figures (6)

  • Figure 1: The flow of our utilized data. We only use one dataset and an iterative approach for data annotation and quality control. The final model is evaluated against the test split. n stands for number of subjects.
  • Figure 2: Showcase of our predicted segmentation masks for random samples from the test set (A -- D). While A and C show a coronal view, B is a sagittal view, and D is an axial one. The rows are, in order, the image, then the image overlayed with the predicted annotation of our final mode, and a 3D view of the segmentation. The colors are the different structures. The air labels are of light colors while the soft tissue ones have the darker shades. Color coding of paranasal sinuses: Red (right maxillary), pink (left maxillary), dark blue (right frontal), light blue (left frontal), orange (right ethmoidal), yellow (left ethmoidal), green (right sphenoidal), turquoise (left sphenoidal).
  • Figure 3: The volume (left) in $mm^3$ and intensity average (right) of different structures across our data. "A." for the aerated/ air-filled space and "ST." for the soft tissue/ pathology proportion.
  • Figure 4: Scatter plots illustrating the relationship between the volume a of A. maxillaris/ frontalis (air-filled or aerated) and ST. maxillaris/ frontalis (soft tissue/pathology). Data points are categorized as "Healthy" (blue) and "Not Healthy" (orange) according to the SHIP radiology reports.
  • Figure 5: Scatter plots illustrating the relationship between the average intensity a of A. maxillaris/ frontalis (air-filled or aerated) and ST. maxillaris/ frontalis (soft tissue/pathology). Data points are categorized as "Healthy" (blue) and "Not Healthy" (orange) according to the SHIP radiology reports.
  • ...and 1 more figures