Dermatofibroma vs. Dermatofibrosarcoma Protuberans: A Nuclear Morphology Study

Abstract

The locally invasive soft tissue sarcoma, dermatofibrosarcoma protuberans (DFSPs), shares certain histologic features of the much more common and benign derssssmatofibroma (DF). While immunohistochemical stains, specifically Cluster of Differentiation (CD) 34 and Factor XIIIa, can be used to distinguish the two entities using microscopy, these markers are not entirely sensitive nor specific.

Three-dimensionally, DFSP nuclei resemble a “puck” or “coin”-like shape. As hematoxylin/eosin-stained slides are prepared, these “puck” nuclei are fixed in an infinite number of orientations depending on their current position in rotation about their axes within the tumor cells. Under histological examination, this random nuclear positioning produces the appearance of two predominate morphologies: an ovoid “disk” shape (en face) and a narrow spindled shape (side-view), which distribute in a roughly 50:50 ratio throughout the tumor sample slide.

Nuclear morphology was analyzed in 324 DFSP and DF samples at high magnification (X 400) to determine the presence or absence of a predominant morphology in which nuclei appear to alternate between an ovoid (en face) and spindled (side-view) throughout the majority of the tumor sample.

An alternating ovoid-spindled nuclear morphology was the predominant cytology in 98% of DFSP and was not predominant in 100% of DF samples (p < .001). This morphology was found to be highly specific (Sp=1) and sensitive (Sn=0.98) for DFSP. This unique nuclear morphology may be a more sensitive and specific diagnostic tool in identifying DFSP from DF in comparison to costly immunohistochemical stains.

Introduction:

Dermatofibrosarcoma protuberans (DFSP) are intermediate grade, locally invasive soft tissue sarcomas. These lesions have a low likelihood of metastasis but possess a high rate of local recurrence, in part due to their propensity to invade deep tissue structures. DFSP account for approximately 18% of all cutaneous soft tissue sarcomas, but overall, are rare lesions, with an incidence rate of 0.8 to 4.5 cases per one million persons per year.^1,2 Particularly in earlier stages, these lesions have a similar appearance to the much more common dermatofibromas (DF), which comprise 3% of all dermatopathology specimens. DFSP’s insidious growth pattern, heterogenous clinical presentation, and close resemblance to benign entities necessitates a high degree of clinical suspicion to avoid delays in diagnosis.

Histologically, DFSP are described as tumor of uniform spindled cells arranged in a storiform pattern, which often invades and entraps subcutaneous fat.³ DFSP are classic translocation sarcomas with uniform cellularity and nuclear morphology throughout, whereas DF typically have variable cellularity, variable nuclear morphology and sclerotic areas with fewer nuclei. In contrast to the characteristic and uniform puck shape of DFSP nuclei, DF nuclei vary in size and shape. The most characteristic demonstrates thin, curved, spindled nuclei arranged in a curlicue pattern around vessels and collagen bundles (sometimes likened to thin black Pacific porpoises jumping over red beach balls). Epithelioid nuclei (vesicular nuclei with prominent central nucleoli) and ringed siderophages are also commonly present in DF.

Immunohistochemical staining is often used to definitively differentiate DFSP from DF. The most frequently used stains are Cluster of Differentiation 34 (CD34), which is positive in DFSP and negative in the central portion of DF, and Factor XIIIa, which is positive in DF and negative in the central portion of DFSP. However, these markers add expense and are not entirely sensitive nor specific; between 6–10% of DFSPs are CD34-negative, and conversely, 15–25% are factor XIIIa positive.^4–6 Other diagnostic markers have been utilized with mixed success, including vimentin, desmin, S100 protein, SMA, keratins, nestin, apolipoprotein D, matrix metalloproteinases (MMP) and cathepsin K.^6–9

Chromosomal translocation t(17;22)(q22;q13), which causes constitutive activation of the tyrosine kinase PDGF-B receptor, is found in 89–96% of DFSPs.⁴ Identification of this translocation using fluorescence in situ hybridization (FISH) may be helpful in cases of recurrent, advanced or inoperable DFSP to assess for utility of imatinib therapy.⁴ Other genetic markers have been identified, providing opportunities for the development of additional molecular target therapies.¹

As hematoxylin/eosin-stained slides are prepared, cell nuclei are fixed in an infinite number of possible orientations depending on their position in space and within the cell at the time of fixation. DFSP nuclei resemble a “puck” or “coin”-like shape. Given this, DFSP fixation results in DFSP nuclei appearing to distribute in a 50:50 ratio of a spindled “side-view” orientation and ovoid “en face” orientation (Figure 1). The objective of this analysis was to describe and compare this microscopic nuclear characteristic within DF and DFSP tumor samples in order to identify additional distinguishing characteristics that could aid in swift diagnosis and reduce cost and burdens associated with special immunohistochemical stains.

Figure 1.

Dermatofibrosarcoma Protuberans (DFSP) Nuclear Morphology.

As slides are prepared, DFSP nuclei are fixed in a 50:50 ratio of spindled “side-view” presentation and “en face” ovoid presentation. The nuclear morphology of dermatofibroma (DF) and DFSP was analyzed at high magnification (X 400) to determine the presence or absence of this morphology.

Materials and Methods:

A retrospective study was performed using archived histological slides. Three hundred thirty-one total hematoxylin/eosin-stained slides (246 DF and 85 DFSP) were evaluated by a single reviewer with regards to pre-determined nuclear morphologic specifications (Table 1). Four DF and three DFSP samples were excluded due to poor slide quality, as determined by the reviewer. Affected cells were compared with adjacent uninvolved epidermis as an internal control for factors such as fixation, processing and staining.¹⁰ Nuclear morphology was analyzed at high magnification (X 400) to determine the presence or absence of a predominant morphology in which nuclei appear to alternate between an ovoid “disk-like” shape (en face) and a narrow spindled shape (side-view) (Figure 1).

Table 1.

Nuclear Morphologic Feature	Grading Schema	Definition
Alternating Ovoid Morphology	Predominant	≥ 75% nuclei in the sample possess an alternating morphology between an ovoid “disk” shape (en face) and a narrow spindled shape (side-view)
	Absent	≤ 75% nuclei in the sample possess this alternating ovoid-spindled morphology

A statistical program IBM^® SPSS^® Version 29 was used for data processing. Differences, proportions, p-values, and positive and negative likelihood ratios [LR (+) and LR (−)] were calculated using 2 × 2 tables, Chi square tests and Fischer-exact test, as appropriate. A LR (+) of more than 10 indicated that alternating ovoid-spindled nuclear morphology had a strong association with DFSP, while 5–10 indicated a moderate association, and less than 5 indicated a weak association with DFSP. A LR (−) of less than 0.1 was strongly associated with DF, 0.1 – 0.5 had a moderate association and more than 0.5 a weak association with DF.¹¹

Results:

The frequencies of a predominant ovoid-spindled nuclear morphology found in DF and DFSP samples is shown in Table 2. Variant tumor characterizations were recorded (Table 3). Of the 82 DFSP included, four samples (4.9%) were classified as variant tumor subtypes: three were characterized as Bednar tumors and one was designated myxoid subtype. Of the 242 DF included, 13 (5.4%) samples contained variant morphologies, which included keloidal, cellular, aneurysmal, glandular, diffuse linear, with monster cells, and with focal osteoid change.

Table 2.

Nuclear Morphology as a Diagnostic Tool to Differentiate DFSP and DF

Nuclear Morphology as a Diagnostic Tool to Differentiate DFSP and DF
	DF (n=242)	DFSP (n=82)	Differences in frequency of features, p-value*
Alternating Ovoid-Spindled Morphology:	0 (0.0%)	80 (97.6%)	<0.001
Sensitivity	0.98
Specificity	1.00
LR+	Infinity**
LR−	0.024**

^*Chi square test

^**Significant diagnostic evidence

Dermatofibroma (DF); Dermatofibrosarcoma Protuberans (DFSP)

Table 3.

Diagnosis	Subtypes	Number of Cases, n	Alternating Ovoid-Spindled Morphology
DF (n=246)	DF unspecified	229	0
	Keloidal	3	0
	Cellular	3	0
	Aneurysmal	2	0
	With monster cells	2	0
	Glandular	1	0
	Diffuse Linear	1	0
	With focal osteoid change	1	0
	Excluded (poor slide quality)	4	Not assessed
DFSP (n=85)	DFSP unspecified	78	77 (98.7%)
	Bednar	3	2 (66.7%)
	Myxoid	1	1 (100%)
	Excluded (poor slide quality)	3	Not assessed

Dermatofibroma (DF); Dermatofibrosarcoma Protuberans (DFSP)

An alternating ovoid- spindled nuclear morphology was the predominant cytology in 98% of DFSP (Figure 2) and was not predominant in 100% of DF samples (p < .001) (Figure 3). Sensitivity, specificity, LR (+) and LR (−) calculations are presented in Table 1. This unique nuclear morphology was found to be highly specific (Sp=1) and sensitive (Sn=0.98) for DFSP.

Figure 2.

Dermatofibrosarcoma Protuberans (DFSP).

High power (x400) showing the bland, uniform storiform pattern of DFSP with alternating ovoid “disk” shape (en face) and narrow spindled shape (side-view) of nuclei.

Figure 3.

High power (x400) of dermatofibroma (DF) showing eosinophilic collagen balls surrounded by plump and pleomorphic fibrohistiocytic cells (collagen entrapment) and lacking the alternating ovoid “disk” shape (en face) and narrow spindled shape (side-view).

Discussion:

Histopathology is often sufficient to diagnose DFSP. Histopathologic descriptions of DFSP include storiform pattern, infiltration into surrounding tissue layers, increased cellularity and occasional mitoses. Identification of these features, in conjunction with the proper gross and clinical picture, can provide strong evidence supporting surgical removal of the lesion.

According to this analysis of 324 samples (242 DF; 82 DFSP), the presence of a distinct alternating nuclear morphology in DFSP is a highly useful diagnostic feature. This morphology, described as alternating between an ovoid (en face) shape and a narrow spindled shape (side view), presents due to the random orientation of the “puck”-like DFSP nuclei rotated about its axis within tumor cells prior to fixation. While DFSP nuclei are relatively uniform in shape within each DFSP sample, their random orientation across numerous tumor cells produces two predominate histological patterns in an approximately 50:50 distribution: ovoid and spindled. This morphology is unique to DFSP, with high specificity and sensitivity. LR(+) and LR(−) values are highly suggestive that this feature provides significant diagnostic evidence.

DF tumors, in contrast to DFSP, possess nuclei of two distinct shapes. These morphologies, histiocytoid and spindled, are each homogenous within individual DF samples and are generally not both found present together within a sample. Histiocytoid-type DF nuclei appear as either polygonal in shape or as elongated epithelioid cells with increased cytoplasm, depending on the nuclear orientation upon fixation. Alternatively, DF with spindled nuclear morphology are more narrow, slender and dark in color. Nuclei throughout a spindled-type DF, likewise, may appear differently within the same sample under microscopy depending on nuclear orientation. DF of both histiocytoid and spindled nuclear morphology did not exhibit the same ovoid-spindled morphology seen in DFSP in the majority of samples. Only two DF samples contained focal areas of resemblance to the DFSP pattern.

In this study, a single reviewer conducted all morphologic analyses within a one-month time-frame in order to maintain internal consistency. Future studies would benefit from a comparison of multiple reviewers’ interpretations of the specimens. Samples used in the analyses were taken from a teaching collection of histologic slides, the compiling of which may have excluded certain samples with more heterogenous or unclear features.

In summary, DFSP tumors exhibit a distinct and predominant alternating ovoid with spindled nuclear morphology. This morphology may be a more sensitive and specific diagnostic tool in identifying DFSP from DF in comparison to costly immunohistochemical stains.