Learn how to systematically approach iMCD pathology.

Histologic Grading of Castleman Disease

Background

Castleman disease (CD) is a cytokine-driven lymphoproliferative disorder with characteristic lymph node histopathology. This standard operating procedure (SOP) provides a method for grading the key histologic features of Castleman disease in excisional lymph node biopsies. It is intended for pathologists, clinicians, trainees and patients as a teaching guide. The procedure can be applied using a conventional light microscope or via digital pathology, using hematoxylin & eosin (H&E) stained slides. Core needle biopsies are not adequate for full histologic evaluation. An intact/excisional lymph node is required for accurate grading of architecture and features.

Tissue and Instrument Requirements

  • Tissue Required: Excisional lymph node biopsy, formalin-fixed-paraffin-embedded (FFPE) and H&E stained. Resection of an entire lymph node is critical; core biopsies lack the architecture to evaluate Castleman feature.
  • Microscopy: Standard light microscope or digital whole-slide image viewer. Ensure the digital viewer can simulate 4x, 10x, 20x, and 40x objectives (approximately 40x, 100x, 200x, 400x total magnification) for proper assessment.
  • Slides: H&E slide of the lymph node (Levels or step-sections can be used if provided, but one good representative section is usually sufficient).
  • Optional: Immunohistochemistry for CD21, CD23, etc., if teaching and illustrating features, as noted above (immunostains are not used for routine grading).

Before grading, scan the entire slide at low magnification to get oriented to the lymph node architecture. Identify where follicles are located and the overall pattern (e.g., nodularity of follicles, interfollicular regions, etc.). Then proceed with the stepwise evaluation below.

01

Low-Power Examination (40x Total Magnification)

At low power (approx. 4x objective, 40x total magnification), assess the overall nodal architecture and follicular features. Three parameters are evaluated at this magnification: follicular twinning, germinal center regression, and hyperplastic germinal centers. Ideally, at least 10 follicles are assessed. These features are best appreciated on a scanning view of the lymph node:

Follicular “Twinning” (Low-power)

Identify if any lymphoid follicles contain two or more germinal centers within a single mantle zone. Twinned germinal centers appear as dual pale centers within one follicle, often enveloped by a shared mantle zone.

Grading (Ideally evaluate at least 10 follicles):
GradeDescription
0No twinned follicles
1~10% to <25% of follicles show twinning
225% to <50%
3≥50% of follicles are twinned

Regressed (Atrophic) Germinal Centers (Low-power)

These are abnormally small, “burned out” germinal centers (approximately <15 lymphocytes across in diameter). They often appear depleted of lymphocytes and may show hyalinization (pink collagenous deposition) in the center. Mantle zones around regressed centers are typically expanded and form concentric rings (“onion-skin” pattern) of small lymphocytes. Sometimes a sclerotic penetrating arteriole is seen coursing through the regressed center at right angle, creating the classic “lollipop” appearance (a vessel with a hyalinized coat in the middle of an atrophic follicle). Regressed germinal centers are a hallmark of the hyaline-vascular (hypervascular) variant of CD.

Grading (Ideally evaluate at least 10 follicles):
GradeDescription
0Germinal centers are normal (0% regressed)
110% to <25% of follicles are regressed
225% to <50%
3≥50% of follicles in the node are regressed/atrophic

Hyperplastic Germinal Centers (Low-power)

These are the opposite of regressed. Hyperplastic GCs are enlarged, hypercellular germinal centers with abundant follicular lymphocytes and active appearance (many tingible body macrophages, high mitotic activity). Follicular hyperplasia is often accompanied by a prominent mantle zone but not onion-skinned (mantle may be thinner or irregular because the germinal center is expanded). This feature is typically prominent in the plasma cell variant of CD or reactive nodes. On low power, hyperplastic GCs appear as numerous large pale nodules (approximately >30 lymphocytes across in diameter).

Grading (Ideally evaluate at least 10 follicles):
GradeDescription
0None or very few small follicles (0% large GCs)
110% to <25% of follicles are large/hyperplastic
225% to <50%
3≥50% of follicles show hyperplasia
Tip: At low power, first count or estimate the total number of follicles in the lymph node section. Then note how many of those show each pattern (twinning, regressed, or hyperplastic). Use the percentage ranges to assign the grade (0-3) for each feature. If using digital slides, you may zoom out to a thumbnail view to help judge overall percentages. Make sure to distinguish regressed vs hyperplastic centers. In some cases of CD, you may see a mix of a few hyperplastic follicles and many regressed ones in the same node (mixed variant). Count each category separately.
02

Medium-Power Examination (100-200x Total Magnification)

Switch to medium power (10x or 20x objective, ~100-200x total magnification) to examine the interfollicular regions in more detail. The key parameter at this magnification is interfollicular vascularity.

Interfollicular Vascularity (Medium-power)

GradeDescription
0Interfollicular vasculature is normal (no obvious increase)
1Mild increase (few more vessels than usual, in 10% to <25% of interfollicular surface area)
2Moderate (definite increase in vessel number in 25% to <50% of surface area)
3Marked hypervascularity (easy to appreciate, ≥50% of interfollicular regions packed with proliferating vessels). A grade 3 vascularity often correlates with easily seeing the piercing vessels at 4x scan (the lollipop sign).
03

High-Power Examination (400x Total Magnification)

Finally, switch to high power (40x objective, 400x total magnification) to assess the interfollicular plasma cells and follicular dendritic cell prominence.

Interfollicular Plasma Cells (High-power)

Evaluate the number of plasma cells in the interfollicular regions. Plasma cells are distinctive with their eccentric nuclei, perinuclear halo, and abundant cytoplasm. Count the number of plasma cells in ideally at least 10 high-power fields (HPFs) across the node.

GradeDescription
0No significant plasmacytosis (only rare plasma cells scattered, essentially 0 to 10% of interfollicular area involved)
1Mild (plasma cells present but sparse – 10% to 25% of cellularity)
2Moderate (notable plasma cells 25% to <50% of interfollicular cellularity)
3Marked plasmacytosis (large confluent sheets of plasma cells replacing ≥50% of interfollicular regions)

Follicular Dendritic Cell (FDC) Prominence (High-power):

Follicular dendritic cells are stromal cells present in germinal centers that form a meshwork to support B-cells and T-cells. In Castleman disease, especially the hyaline-vascular type, FDCs can be prominent and dysplastic in the regressed germinal centers. A regressed GC may consist almost entirely of FDCs with very few lymphocytes. At high power, FDC prominence is recognized by the presence of numerous spindle to ovoid nuclei in the germinal center and a pale eosinophilic stroma in GC. Immunostains like CD21 can highlight this meshwork, but are not a part of grading.

GradeDescription
0FDCs are not particularly conspicuous (germinal centers have normal cellular makeup, you cannot readily identify FDCs on H&E)
1Slight prominence (<25% of cellularity of GC)
2Moderate (FDC meshwork obvious in many follicles 25% to <50%)
3Marked FDC prominence (≥50% of cells are FDCs)

Use of Light Microscopy vs Digital Pathology

This grading SOP can be executed via traditional microscopy or digital slides:

  • When using light microscopy, ensure the microscope is calibrated and use the appropriate objective for each step (4x for low power, 10x/20x for medium, 40x for high). A multi-head microscope session can be useful for consensus grading or teaching, allowing multiple observers to discuss what constitutes each grade.
  • When using digital pathology, verify that the scanning magnification (ie., 40x scan) provides sufficient resolution. Use the viewer’s magnification indicators to approximate 4x, 10x, 20x, 40x views. Digital viewing makes it easy to navigate at low power; however, be cautious at high power to distinguish plasma cells and FDC details. Digital annotations can be helpful (for example, outlining a follicle that is regressed vs one that is hyperplastic for teaching or consultation).

Summary

For convenience, the below summarizes the histologic features, definitions, and scoring criteria:


FeatureDescriptionGrade 0Grade 1Grade 2Grade 3
Follicular Twinning>1 germinal center in one follicle
(shared mantle zone, low power)
None
(0%)
~10% to <25%
follicles twinned
25% to <50%≥50%
follicles twinned
Regressed Germinal CentersAtrophic, lymphocyte-depleted GCs
± hyaline, onion-skin mantle
None
(0%)
10% to <25%
follicles regressed
25% to <50%≥50%
follicles regressed
Hyperplastic Germinal CentersEnlarged, reactive-appearing GCsNone or very few
(0%)
10% to <25%
follicles hyperplastic
25% to <50%≥50%
follicles hyperplastic
Interfollicular VascularityProliferation of blood vessels
in interfollicular areas
NormalMild
(10% to <25%)
Moderate
(25% to <50%)
Marked
(≥50%)
PlasmacytosisExcess plasma cells
in interfollicular zones
None
(0% to 10%)
Mild
(10% to 25%)
Moderate
(25% to <50%)
Marked
(≥50%)
FDC ProminenceExpansion of follicular dendritic cell networks
(often filling regressed GCs)
Normal meshworkSlight
(<25%)
Moderate
(25% to <50%)
Marked
(≥50%)

Each feature should be independently scored as grade 0, 1, 2, or 3 as defined. Do not sum the scores to make a diagnosis. Current diagnostic criteria for Castleman disease do not use a total score cutoff. Instead, the presence of high-grade changes in either spectrum (vascular/FDC or plasma cell) helps classify the histopathological variant. In fact, the World Health Organization (WHO) classification for idiopathic multicentric CD (iMCD) requires at least a grade 2 or 3 in either regressed GCs or plasmacytosis (or both), along with supportive clinical findings, to diagnose CD.


Reporting

Remember that no single histologic feature is pathognomonic for Castleman disease: all features must be interpreted together. As research progresses, the hope is to quantify these changes for prognostic or diagnostic indices, but at present the grading is primarily for teaching and documentation. Always mention in the bottom line or comment of your report that an excisional biopsy was used and note the dominant histologic pattern and grade (ie, “Castleman disease, hypervascular type, with marked regressed germinal centers (grade 3), hypervascularity (grade 3), and minimal plasmacytosis (grade 0-1)”). This conveys the key findings to clinicians.


References

  1. Fajgenbaum et al., Blood 2017; iMCD
  2. Alnoor et al., IJLH 2024

Finished the courses? Explore our image library for independent study.

Explore Library