Optical Diagnosis of Dermatofibrosarcoma Protuberans Differentiated from Dermatofibroma Using Non-Linear Optical Microscopy
Dermatofibrosarcoma protuberans (DFSP) and dermatofibroma (DF) are cutaneous lesions with overlapping histological features, often leading to diagnostic challenges. This study demonstrates the application of label-free non-linear optical (NLO) microscopy as a powerful tool to differentiate DFSP from DF by analyzing intrinsic tissue components such as collagen, elastin, and cellular structures. The methodology leverages second-harmonic generation (SHG) and two-photon excited fluorescence (TPEF) signals to provide high-resolution, quantitative insights into the microstructural and biochemical differences between these lesions.
Tissue Microstructure and Collagen Distribution
NLO imaging revealed distinct architectural differences between normal skin, DF, and DFSP. In normal skin, collagen fibers exhibited a gradient of organization, appearing thinner in the superficial dermis (region of interest [ROI] 1) and progressively coarser in the middle (ROI 2) and deep dermis (ROI 3). SHG signals, generated by polarizable collagen bundles, dominated the dermal layers, with intensities reflecting collagen density. Elastic fibers, detected via TPEF, were uniformly distributed across the dermal layers at low densities (3.3 ± 0.7% in ROI 1, 4.3 ± 2.8% in ROI 2, and 2.4 ± 1.1% in ROI 3).
In DF lesions, NLO images showed a hyperplastic epidermis with basal layer hyperpigmentation. The tumor penetrated only the superficial and middle dermis (ROIs 1 and 2), sparing the deep dermis (ROI 3), where collagen organization resembled normal skin. Collagen density in DF was significantly reduced in ROIs 1 (14.1 ± 5.8%) and 2 (23.2 ± 3.3%) compared to normal skin (42.6 ± 6.7% and 46.5 ± 5.2%, respectively; P < 0.001). Elastic fibers accumulated in ROI 1 (10.3 ± 4.6%) but were nearly absent in ROI 2 (0.2 ± 0.2%). The lesion margin was smooth, with thickened collagen bundles and a mixture of stromal cells, histiocytes, and blood vessels.
In contrast, DFSP displayed an atrophic epidermis overlying a narrow tumor-free Grenz zone. The tumor infiltrated deeper into the subcutaneous tissue, forming a honeycomb pattern with spindle cells oriented parallel to the skin surface. SHG signals were markedly diminished in ROIs 2 and 3 (collagen density: 2.7 ± 2.1% and 0.4 ± 0.2%, respectively), indicating non-polarizable collagen. TPEF signals from elastin were elevated in ROI 1 (15.0 ± 3.0%) and ROI 2 (13.6 ± 4.8%), forming a coarse network in the middle dermis.
Spectral Analysis and SHG-to-TPEF Index
NLO spectra normalized to SHG (405 nm) and TPEF (475–535 nm) signals provided further diagnostic criteria. In normal skin, SHG signals from collagen dominated all dermal layers. DF exhibited stronger TPEF in ROI 1 due to elastin accumulation, while DFSP showed suppressed SHG signals in ROIs 2 and 3, with TPEF surpassing SHG intensity.
The SHG-to-TPEF index (STI), calculated as (SHG − TPEF)/(SHG + TPEF), quantified these differences. STI values were positive in normal skin and DF (indicating SHG > TPEF) but negative in DFSP for ROIs 2 (−0.44 ± 0.21) and 3 (−0.42 ± 0.22), confirming TPEF dominance. This inversion directly correlated with the loss of polarizable collagen in DFSP, offering a clear diagnostic marker.
Quantitative Comparison of Matrix Components
Collagen and elastin densities were quantified by pixel analysis of SHG and TPEF signals. In DF, collagen density in ROI 3 (37.6 ± 4.6%) matched normal skin (37.8 ± 7.2%; P = 0.962), confirming limited invasion. DFSP, however, showed near-zero collagen density in ROIs 2 and 3, alongside disrupted superficial dermal collagen (21.2 ± 8.1% in ROI 1).
Elastin distribution also differed: DFSP had elevated elastin in ROI 2 (13.6 ± 4.8%) compared to normal skin (4.3 ± 2.8%; P < 0.001), while DF showed minimal elastin in ROI 2 (0.2 ± 0.2%). These patterns highlighted elastin reorganization as a secondary diagnostic feature.
Clinical Implications and Advantages of NLO Microscopy
Traditional histopathology relies on H&E staining, which may obscure subtle differences in collagen organization. NLO microscopy bypasses staining artifacts, enabling direct visualization of collagen polarity and elastin distribution. The method’s quantifiable parameters, such as STI and matrix densities, reduce subjectivity in diagnosis.
For DF, the preservation of polarizable collagen in the deep dermis and elastin accumulation in the superficial layer distinguish it from DFSP. In DFSP, the absence of SHG signals in deeper layers, combined with elastin network formation in the middle dermis, provides a definitive signature. These findings align with histopathological features—such as storiform spindle cells in DFSP—but add objective metrics to support differential diagnosis.
Conclusion
This study establishes NLO microscopy as a robust, label-free alternative for distinguishing DFSP from DF. By quantifying SHG and TPEF signals, the technique identifies collagen disorganization, elastin redistribution, and cellular infiltration patterns unique to each lesion. The approach enhances diagnostic accuracy, particularly in borderline cases, and paves the way for automated, image-based diagnostic systems. Future studies with larger cohorts could further validate these biomarkers and refine diagnostic thresholds.
doi.org/10.1097/CM9.0000000000001548
Was this helpful?
0 / 0