Using Imaris to rigorously track PET-defined sites of lung inflammation in Mycobacterium tuberculosis-exposed non-human primates

This paper proposes and validates a novel Imaris-based pipeline for rigorously tracking and characterizing 3D PET-defined lung inflammation sites in *Mycobacterium tuberculosis*-exposed non-human primates, offering superior alignment, automated segmentation, and detailed morphological metrics compared to traditional 2D slice-based analysis methods.

The Gist

Imagine your lungs are a vast, foggy forest. When a non-human primate (like a monkey) breathes in the bacteria that causes tuberculosis (Mycobacterium tuberculosis), it doesn't infect the whole forest at once. Instead, it sets up tiny, hidden "camps" or clearings where the immune system starts fighting back. These camps are invisible to the naked eye, but they glow like little bonfires when viewed through a special camera called a PET/CT scanner, which uses a radioactive sugar to light up active inflammation.

The Old Way: Counting Bonfires with a Ruler
Traditionally, scientists have looked at these glowing scans like they are looking at a stack of flat paper maps (2D slices). They manually point at each glowing camp on every single slice, trying to guess the size and brightness of the fire. It's a bit like trying to measure the volume of a cloud by tracing its outline on a piece of paper, slice by slice. While this gives them a rough idea of how bright the fire is (called the "SUV"), it's tedious, prone to human error, and doesn't give a true 3D picture of the camp's shape or size.

The New Way: Building a 3D Hologram
This paper introduces a clever new method using software called Imaris. You might think of Imaris as a tool usually reserved for biologists studying tiny cells under a microscope, but the researchers realized it's perfect for this job too.

Here is how their new "pipeline" works, using a few simple analogies:

1. The Anchor Points (Landmarks):
   Imagine you are trying to line up a series of photos of the same person taken over several years. If the person moves, the photos won't match up perfectly. To fix this, the researchers use the monkey's spine vertebrae (the bones in its back) as "anchors" or landmarks. Just like you might use a person's nose and ears to align a photo, the software uses the spine to perfectly stack all the different scans on top of each other, ensuring they are looking at the exact same spot in the body every time.

2. The Digital Clay (3D Segmentation):
   Instead of tracing flat slices, the software treats the glowing inflammation like a lump of digital clay. It automatically molds a 3D "skin" or "surface" around the glowing fire. If the inflammation is a perfect sphere, the software wraps a sphere around it. If it's a weird, jagged shape, the software molds the clay to fit that exact shape. This creates a true 3D hologram of every single infection site in the lungs and the nearby lymph nodes (the body's security checkpoints).

3. The Double-Check:
   The researchers tested this new method against the old one. They found that the brightness measurement from the new 3D method matched the old 2D method almost perfectly. This proved that their new way is just as accurate for measuring "how hot the fire is," but much better at everything else.

Why This Matters: The "Virtual Reality" Bonus
The real magic of this new approach is the extra data it unlocks. Because the software builds a full 3D model, it can tell you:

• Volume: Exactly how much space the infection takes up.
• Shape: Is it a smooth ball or a spiky star?
• Location: Precisely where it sits in the lung.

But the coolest part? The researchers can export these 3D models into a Virtual Reality file (.wrl). Imagine putting on a VR headset and being able to "walk" inside the monkey's lung, floating around the infection sites, rotating them, and examining them from every angle.

The Bottom Line
This paper is about upgrading from a flat, 2D sketch of a disease to a fully immersive, 3D holographic model. By using a tool designed for tiny cells to look at whole lungs, and by using the spine as a guide to keep everything aligned, scientists can now track how tuberculosis infections grow, shrink, or change shape over time with incredible precision. This helps them understand if a new treatment is working much better than ever before.

Technical Summary

Technical Summary: Using Imaris for PET-Defined Lung Inflammation Tracking in Mtb-Exposed NHPs

1. Problem Statement
In studies involving non-human primates (NHPs) exposed to Mycobacterium tuberculosis (Mtb) via aerosol, inflammation manifests as discrete sites within the lungs. These sites are typically detected using 2-deoxy-2-[fluorine-18]fluoro-D-glucose (18F-FDG) PET/CT scans. However, current analytical workflows rely heavily on software like Invicro VivoQuant or OsiriX, which present significant limitations:

• Manual Labor: Analysis requires manual labeling of PET signal sites using 2D slices, which is time-consuming and prone to inter-observer variability.
• Limited Metrics: Reporting is often restricted to the Maximum Standardized Uptake Value (SUV) of the whole lung or individual lesions, lacking comprehensive morphological data.
• Lack of 3D Integration: The approach does not fully leverage 3D spatial relationships or facilitate rigorous longitudinal tracking of lesion evolution.

2. Methodology
The authors propose a novel analytical pipeline utilizing Imaris, a proprietary software traditionally designed for fluorescent microscopy, adapted here for 3D PET/CT data analysis. The workflow includes:

• Rigid Registration via Landmarks: To align serial scans of the same animal over time, the system utilizes the anatomical positions of spine vertebrae as fixed "landmarks" for precise spatial registration.
• 3D Automated Segmentation: Instead of 2D slicing, the pipeline employs automated image segmentation to generate "surfaces" representing 3D volumes. This process is refined with manual corrections to ensure accuracy.
• Target Identification: The segmentation specifically isolates sites of inflammation within the lung parenchyma and lung-associated thoracic lymph nodes (LNs).
• Data Export: Identified lesions are exported in Virtual Reality file formats (.wrl) to facilitate detailed 3D visualization and further analysis.

3. Key Contributions

• Software Adaptation: Successfully repurposing Imaris, a microscopy tool, for high-resolution PET/CT analysis in a preclinical NHP model.
• Automated 3D Workflow: Shifting from manual 2D slice labeling to an automated 3D surface segmentation approach, significantly enhancing the precision of lesion definition.
• Comprehensive Morphometrics: Moving beyond simple intensity metrics to capture a "wealth" of additional lesion features, including:
  • Volume
  • Spatial location
  • Shape parameters
  • Surface area
• Longitudinal Tracking: Establishing a robust method to track the evolution of specific lesions over time and correlate these changes with infection outcomes or treatment efficacy.

4. Results

• Validation: The study demonstrated an excellent correlation between the Maximum SUV values of individual lesions calculated by the traditional Invicro VivoQuant software and the Maximum Intensity values derived from the new Imaris pipeline. This validates the accuracy of the Imaris approach for intensity-based quantification.
• Enhanced Resolution: The pipeline successfully defined the location of all inflammation sites, including those in the lungs and thoracic lymph nodes, with high spatial fidelity.
• Data Richness: The method generated multidimensional data for each lesion, allowing for complex analyses of how lesion morphology (shape, volume, surface area) changes dynamically during the course of infection and treatment.

5. Significance
This research represents a paradigm shift in the quantitative analysis of Mtb infection in NHP models. By transitioning from 2D manual labeling to 3D automated segmentation, the study offers:

• Rigor and Reproducibility: Reducing human error and increasing the consistency of lesion tracking across time points.
• Deeper Biological Insight: Enabling researchers to correlate not just metabolic activity (SUV), but also physical lesion characteristics (volume, shape) with clinical outcomes.
• Future-Ready Analysis: The ability to export lesions in VR formats (.wrl) opens new avenues for immersive, detailed examination of disease progression, potentially improving the design and evaluation of tuberculosis vaccines and therapeutics.