ImageJ Fibrosis Quantification (MT, PSR, IHC) 2026 Guide

Introduction: Why Quantitative Image Analysis Matters

In fibrosis research, biochemical bulk assays such as hydroxyproline quantification and ELISA are necessary but insufficient on their own. Spatial and morphological information—whether fibrosis is perivascular, interstitial, or accompanied by architectural remodeling—can only be demonstrated quantitatively through image analysis.

Traditional pathology assessment relies on semi-quantitative scoring systems (e.g., 0–4 scales), which are inherently subjective and suffer from poor inter-observer reproducibility. ImageJ (Fiji distribution) has emerged as the standard open-source tool for computing area fraction (%Area), now widely adopted in both peer-reviewed publications and preclinical study reports¹.

This article provides hands-on protocols for analyzing three major staining methods in ImageJ: Masson's trichrome (MT), picrosirius red (PSR), and immunohistochemistry (IHC/DAB).

---

Quick Reference: Workflow Summary

Stain	Color Separation Method	Threshold Target	Recommended Auto-Threshold	Output Metric
Masson's trichrome (MT)	Color Deconvolution (Masson Trichrome vector)	Blue component	Huang / Otsu	%Area (collagen area fraction)
Picrosirius red (PSR), brightfield	RGB Split → Green channel	Red component (inverted)	Otsu	%Area
PSR, polarized light	8-bit conversion	Luminance	Default / Triangle	%Area (birefringence area fraction)
IHC / DAB	Color Deconvolution (H DAB vector)	DAB component	MaxEntropy / Otsu	%Area (positive area fraction)

Key principle: Within any given project, use the same thresholding algorithm and parameter values across all images to minimize operator bias.

---

1. ImageJ/Fiji Setup and Recommended Plugins

1-1. Installing Fiji

Fiji (Fiji Is Just ImageJ) is a batteries-included distribution of ImageJ bundled with curated plugins for life-science image analysis¹. Download the appropriate version for your operating system from the official site (https://imagej.net/software/fiji/).

1-2. Recommended Plugins

Plugin	Purpose	Bundled with Fiji?
Color Deconvolution	Optical density separation by stain chromogen	Yes
Auto Threshold	Side-by-side comparison of multiple thresholding algorithms	Yes
Bio-Formats	Import whole-slide images (NDPI, SVS, etc.)	Yes
StarDist / Cellpose	Deep-learning cell segmentation (advanced workflows)	Manual install

1-3. Image Acquisition Best Practices

• White balance: Calibrate white balance at the time of capture and keep settings consistent across all slides in a study. Post-hoc correction introduces artifacts into quantitative measurements.
• Consistent resolution: Maintain uniform magnification and pixel size within a project. Calibrate using Analyze > Set Scale referencing the scale bar.
• File format: Use uncompressed TIFF whenever possible. JPEG compression introduces block artifacts that degrade Color Deconvolution accuracy.

---

2. Protocol A: Masson's Trichrome (MT) Quantification

In Masson's trichrome staining, collagen fibers stain blue (or green in some variants), muscle and cytoplasm stain red, and nuclei stain dark purple. Thresholding a specific color directly from a full RGB image is unreliable, making Color Deconvolution the essential preprocessing step².

Step 1: Color Deconvolution

1. Open the image and navigate to Image > Color > Color Deconvolution.
2. From the Vectors dropdown, select "Masson Trichrome".
   • If staining conditions differ from the built-in vectors (e.g., different dye lots or scanner profiles), select three ROIs representing blue stain, red stain, and background, then create custom vectors via From ROI for improved separation accuracy.
3. Three grayscale images (Colour_1, 2, 3) will be generated. Retain only the image where collagen (blue component) appears darkest, and close the others.

Step 2: Thresholding

1. Select the target image and open Image > Adjust > Threshold (Ctrl+Shift+T).
2. Adjust the sliders so that collagen regions are highlighted in red. Review the histogram to confirm appropriate separation from the background.
   • Test automatic thresholds (Huang, Otsu, etc.) and apply a consistent algorithm or fixed numerical value across all images in the project.
3. Click Apply to generate a binary (black-and-white) image.

Step 3: Area Fraction Measurement

1. In Analyze > Set Measurements, check "Area", "Area Fraction", and "Limit to threshold".
2. Run Analyze > Measure (Ctrl+M).
3. The %Area value in the Results window represents the collagen area fraction relative to the total image area.

For a detailed comparison of sensitivity and specificity between trichrome and Sirius red staining, see Masson's Trichrome vs. Sirius Red.

---

3. Protocol B: Picrosirius Red (PSR) Quantification

As detailed in Sirius Red Staining: Protocol and Principles, PSR stains collagen fibers red. Two distinct observation modes—brightfield and polarized light—each call for a different analytical approach.

3-1. Brightfield Image Quantification (RGB Split)

1. Open the image and run Image > Color > RGB Split.
2. Select the Green channel (red-stained collagen shows the highest contrast in the green channel).
3. Apply Image > Adjust > Threshold to binarize collagen regions, then run Measure to calculate %Area.

3-2. Polarized Light Image Quantification

Under polarized light microscopy, collagen exhibits bright birefringence against a dark background, making it particularly amenable to automated quantification.

1. Open the image and convert to grayscale via Image > Type > 8-bit.
2. Apply Image > Adjust > Threshold to exclude background noise and highlight birefringent collagen only.
3. Run Measure to calculate %Area.

3-3. Collagen Subtype Quantification (HSB Analysis)

To separately quantify Type I collagen (thick fibers, red–yellow birefringence) and Type III collagen (thin fibers, green birefringence):

1. Open the polarized light image and convert via Image > Type > HSB Stack.
2. Select the Hue channel and set hue-range thresholds (e.g., red–yellow = 0–40; green = 80–130).
3. Measure %Area for each hue range independently.

See also PSR Staining Troubleshooting for common pitfalls and solutions.

---

4. Protocol C: Immunohistochemistry (IHC/DAB) Quantification

This protocol applies to tissue sections stained with DAB chromogen (brown) and hematoxylin counterstain (blue), commonly used for markers such as alpha-SMA, collagen type I, and various macrophage subtype markers.

Step 1: Color Deconvolution (H DAB Vector)

1. Open the image and run Image > Color > Color Deconvolution, selecting "H DAB" from the Vectors dropdown².
2. From the three output images, select the DAB (brown) channel (typically Colour_2).

Step 2: Thresholding and Measurement

1. Apply Threshold to the DAB channel to binarize the positive signal.
2. Run Measure to obtain the DAB-positive area fraction (%Area).

Tip: To calculate the H-score (staining intensity x positive area), avoid simple binarization. Instead, use the intensity histogram to categorize regions into weak, moderate, and strong staining, then measure the area of each intensity category separately.

---

5. Batch Processing with Macros

Manually analyzing dozens or hundreds of images is both time-consuming and error-prone. ImageJ's macro language enables full automation of the workflow.

5-1. Recording a Macro

Open Plugins > Macros > Record, then perform the protocol steps on a single image. The Recorder window will generate an IJM script capturing each operation.

5-2. Example Batch Macro (MT Staining)

// Masson's Trichrome Batch Quantification Macro
inputDir = getDirectory("Input folder");
outputDir = getDirectory("Output folder");
list = getFileList(inputDir);

for (i = 0; i < list.length; i++) {
    if (endsWith(list[i], ".tif") || endsWith(list[i], ".jpg")) {
        open(inputDir + list[i]);
        id = getImageID();
        run("Colour Deconvolution", "vectors=[Masson Trichrome]");
        // Select the blue channel (Colour_3)
        selectWindow(list[i] + "-(Colour_3)");
        setAutoThreshold("Huang dark");
        run("Set Measurements...",
            "area area_fraction limit redirect=None decimal=3");
        run("Measure");
        close("*");
    }
}
saveAs("Results", outputDir + "MT_quantification.csv");

Note: The window name in selectWindow may vary depending on the Color Deconvolution plugin version. Use the Record function to verify the actual window name before incorporating it into your macro.

5-3. Running the Macro

Paste the code into Process > Batch > Macro, or save it as an .ijm file and run via Plugins > Macros > Run.

---

6. Troubleshooting

Threshold Variability

Problem	Cause	Solution
Large %Area variation across samples	Staining intensity varies between batches	Stain all sections in a single batch. Create custom vectors via From ROI
Auto-threshold fails on certain images	Low tissue content or large blank areas	Define an ROI encompassing tissue only before measurement
Non-collagen structures scored as positive	Vessel walls, air bubbles, or tissue folds	Manually exclude artifacts with ROIs or generate exclusion masks

White Balance Issues

• Poor white balance at acquisition directly compromises Color Deconvolution separation accuracy.
• Verify that blank regions (glass-only areas) have RGB values of at least (220, 220, 220).
• If post-hoc correction is unavoidable, consider Process > Subtract Background (Rolling Ball), but validate the impact on quantitative values before applying to study data.

Managing Batch Effects

• Images acquired on different scanners or on different days carry systematic bias.
• Include control slides in every staining batch and confirm that control %Area values fall within an acceptable range before analyzing experimental data.
• Consider including batch as a covariate in statistical analysis to account for residual variability.

---

7. Where Image Quantification Fits in Fibrosis Assessment

ImageJ-based area fraction analysis provides morphological and spatial quantitative data within the broader toolkit of fibrosis evaluation approaches described in the Comprehensive Fibrosis Assessment Guide—including histological scoring, biochemical assays, and gene expression profiling. Combining hydroxyproline-based total collagen measurements with ImageJ-derived localization and distribution data (%Area) substantially strengthens the evidence base for drug efficacy evaluation in preclinical studies.

While AI-powered digital pathology platforms (deep learning-based segmentation) are gaining traction, ImageJ/Fiji remains a foundational tool: it is open-source, highly reproducible, and its methodology is fully transparent and verifiable by peer reviewers and regulatory agencies alike.

---

FAQ

Q1. IHC %Area vs PSR %Area — which should be the first-line readout?

IHC quantifies specific molecular species (α-SMA, collagen type I/III), whereas PSR captures total collagen as a generic fibrosis proxy. If your drug mechanism targets a specific protein, IHC is the right readout; for tracking total fibrotic burden, PSR is the first-line choice. The two methods are complementary rather than substitutes, and preclinical studies typically include both. For method selection criteria, see the MT vs PSR comparison guide and the fibrosis quantification comparison framework.

Q2. Should ImageJ %Area and hydroxyproline assay be used together?

Yes — dual readout combining histology (spatial distribution) and biochemistry (absolute quantification) is the best practice in preclinical fibrosis studies. Typical correlation is r = 0.7–0.85; discordance usually reflects heterogeneous fibrosis distribution, which would be misread if either method were used alone. Reported CV% runs 15–25% for PSR %Area versus 5–10% for hydroxyproline, so regulatory submissions gain persuasive power when both are reported side-by-side. See the fibrosis quantification comparison framework for a full breakdown.

Q3. When should I migrate from ImageJ to AI pathology (QuPath / deep learning)?

Three pragmatic triggers: (1) slide count — AI becomes cost-effective above ~100 slides per project; (2) inter-observer CV — if manual re-analysis by multiple operators yields > 15% variation, AI delivers stronger reproducibility; (3) GLP / regulatory requirements — AI deployments require additional validation documentation for audit. ImageJ retains advantages in transparency and reviewer verifiability, so continuing with it remains rational for smaller studies. See the AI pathology fibrosis guide for the migration checklist.

Q4. Are Ashcroft score and ImageJ %Area both necessary?

They address different questions and are usually combined in preclinical studies. Ashcroft is a semi-quantitative "structural" index capturing lesion distribution (focal vs diffuse); ImageJ %Area is a continuous "quantitative" index. In FDA/PMDA submissions, %Area is commonly assigned as the primary endpoint (continuous variables deliver higher statistical power), with Ashcroft as a secondary endpoint (interpretable to clinical pathologists). See the Ashcroft score interpretation guide for scoring rules.

Q5. Masson's Trichrome vs Picrosirius Red — which is better suited to ImageJ quantification?

PSR generally yields higher quantitative precision. Three reasons: (i) the red PSR signal shows high contrast in the green channel, giving stable auto-thresholding; (ii) under polarized light, background noise drops dramatically, improving CV% by 5–10 percentage points; and (iii) MT's blue component tends to bleed into the faint blue signal of nuclei and muscle, causing misclassification during Color Deconvolution. Conversely, MT retains value when multi-component assessment (fibrin, muscle, cytoplasm) is required. See the MT vs PSR comparison guide.

Q6. Whole Slide Image (NDPI/SVS) vs ROI-based quantification — which should I adopt?

WSI is ideal for representativeness (eliminates field-of-view selection bias). However, ImageJ + Bio-Formats is RAM-bound, with a practical ceiling of ~1–2 GB per slide, making whole-slide analysis impractical for large cohorts. The pragmatic compromise is a staged approach: use ROI sampling (5–10 fields per slide) at the screening stage, then switch to full WSI quantification (QuPath or AI-integrated platforms are the de facto standard) for pivotal confirmatory studies. At the WSI stage, integration with AI pathology platforms is usually required.

---

Related Articles

• Masson's Trichrome Staining Protocol Guide — MT staining procedures and quantification tips
• PSR Staining Troubleshooting Guide — Five common PSR staining problems and fixes
• Multiplex Immunofluorescence (mIF) Protocol — Multi-marker spatial analysis for fibrotic tissues
• Masson's Trichrome vs Sirius Red Comparison — Side-by-side comparison of principles, CV, and cost
• Comprehensive Fibrosis Assessment Guide — Overview of fibrosis evaluation methods

---

References

1. Schindelin J, et al. Fiji: an open-source platform for biological-image analysis. Nat Methods. 2012;9(7):676-682. (PubMed)

2. Ruifrok AC, Johnston DA. Quantification of histochemical staining by color deconvolution. Anal Quant Cytol Histol. 2001;23(4):291-299. (PubMed)

3. Lattouf R, et al. Picrosirius red staining: a useful tool to appraise collagen networks in normal and pathological tissues. J Histochem Cytochem. 2014;62(10):751-767. (PubMed)

4. Varghese F, et al. IHC Profiler: an open source plugin for the quantitative evaluation and automated scoring of immunohistochemistry images of human tissue samples. PLoS One. 2014;9(5):e96801. (PubMed)

5. Chen Y, et al. Quantification of collagen in liver fibrosis: Sirius Red staining should be a standard method. World J Gastroenterol. 2021;27(30):4952-4968. (PubMed)