The correlation of pigment content between dermoscopy and histopathology in basal cell carcinoma: A retrospective study

Introduction

Basal cell carcinoma (BCC), the most common human skin cancer, has shown a rising global incidence, imposing a significant burden on patients and the healthcare system.^1,2 It often exhibits varying degrees of pigmentation, complicating differentiation from melanoma and benign pigmented skin lesions, potentially leading to misdiagnosis or excessive intervention.^1,3,4

The advent of dermoscopy has revolutionised BCC diagnosis, elevating accuracy from approximately 60% to over 90% by enabling the visualisation of pigmented and vascular structures.^5,6 This technique is particularly advantageous for identifying pigmented structures and delineating tumour margins.^7,8

Notably, pigmentation occurs in nearly all BCC subtypes (nodular, superficial and infiltrating types), while dermoscopy reveals pigmented structures in 30% of clinically non-pigmented cases.^9,10 Histopathologically, pigmentation at the dermal-epidermal junction appears brown (e.g., leaf-like areas, spoke wheel-like structures, and dots), while deeper dermal melanin appears blue/grey (e.g., blue-grey ovoid nests/globules).^11,12 Asian studies have indicated that pigmentation may inversely correlate with tumour aggressiveness, with less pigmented BCCs showing higher invasiveness.^13,14 Additionally, dermoscopic features (e.g., pigmented structures, ulceration, arborising telangiectasias) predict treatment response and tumour persistence.^15-17

However, dermoscopy optically reflects tissue pigmentation with a poorly quantified correlation to histopathology. Most prior studies focused on structural rather than pigment content, leaving a gap in understanding on how dermoscopic pigmentation mirrors histopathological melanin distribution. Given high BCC prevalence and pigmentation in Asian populations, this study retrospectively analysed the correlation between dermoscopy and histopathology with respect to pigment content.

Methods

Patients and images

Clinical/dermoscopic images of histopathologically confirmed BCCs were extracted from the dermatology outpatient departments of China-Japan Friendship Hospital between September 2018 and June 2022. The clinical images were obtained using standard digital photography, while dermoscopic images were acquired with a 20x magnification digital dermoscopy system (Medicam 800HD, FotoFinder Systems GmbH, Birbach, Germany). Cases with thick scales/scabs or suboptimal images were excluded. The study was approved by the Ethics Review Committee and conducted in accordance with the Declaration of Helsinki. Informed consent was obtained from all patients or their legal guardians. All dermoscopic diagnoses and characterisations were described using International Dermoscopic Association terminology by two physicians blinded to histopathology. Histopathological diagnosis was confirmed by two associate chief dermatopathologists. All tumours were surgically excised without adjuvant therapy. The flow chart of our research has been illustrated in Figure 1.

Close

Figure 1:

Research flow chart.

Export to PPT

Dermoscopic and histopathological assessment

In the absence of a standardised pigment grading system, lesions were classified into four pigment categories based on visual estimation [Figure 2]: (i) non-pigmented (no visible pigmentation); (ii) light pigment (pigmentation involving <30% of lesion area); (iii) moderate pigment (pigmentation involving 30% to 70% of lesion area), and (iv) heavy pigment (pigmentation involving >70% of lesion area).²² Grading was performed independently by two observers, with discrepancies resolved through consensus.

Close

Figure 2a-c:

(a) Clinical photograph of BCC located on the nasal alar region of Patient A, (b) The non-pigment under non-polarised dermoscopy in BCC showing ulceration and arborizing vessels (20×), (c) No pigment under histopathology in BCC, stained with Haematoxylin and eosin staining, 20x.

Export to PPT

Close

Figure 2d-f:

(d) Clinical photograph of BCC located on the nasal alar region of Patient B, (e) The light pigment under non-polarised dermoscopy in BCC (yellow arrows and red boxes) showing blue white veil, ulceration and blue-grey ovoid nests (20×), (f) No pigment under histopathology in BCC, stained with Haematoxylin and eosin, 20x.

Export to PPT

Close

Figure 2g-i:

(g) Clinical photograph of BCC located on the nasal alar region of Patient C, (h) The moderate pigment under non-polarised dermoscopy in BCC (yellow arrows and red boxes) showing leaf-like areas, arborising vessels and blue-grey globules (20×), (i) Light pigment under histopathology in BCC (yellow arrows and red boxes), stained with Haematoxylin and eosin, 20x.

Export to PPT

Close

Figure 2j-l:

(j) Clinical photograph of BCC located on the nasal tip of Patient D, (k) The heavy pigment under non-polarised dermoscopy in BCC (yellow arrows and red boxes) showing leaf-like areas, brown dots and blue white veil (20×), (l) Light pigment under histopathology in BCC (yellow arrows and red boxes), stained with Haematoxylin and eosin, 20x.

Export to PPT

Results

The connection of pigment content with dermoscopy and histopathology

The dermoscopic examination showed that only a few BCCs were non-pigmented (3/90, 3.3%). BCCs with light, moderate, and heavy pigment accounted for 15.6% (14/90), 32.2% (29/90), and 48.9% (44/90) of cases, respectively. According to the histopathological diagnosis, 30 BCCs (33.3%) were non-pigmented. Light, moderate, and heavy pigmented BCCs accounted for 53.3% (48/90), 12.2% (11/90), and 1.1% (1/90) of cases, respectively. A moderate correlation was observed between dermoscopic and histopathological pigment grades, with a correlation coefficient of 0.5028 (p<0.05) [Figure 3]. Interobserver agreement for pigment grading was poor (Cohen’s kappa = −0.032, p > 0.05), necessitating the use of consensus grading for final evaluation.

Close

Figure 3:

Correlation analysis of dermoscopic pigment content and histopathologic pigment content in BCC.

Export to PPT

Discussion

Our study retrospectively quantified the correlation between dermoscopic and histopathological pigmentation in 90 BCCs and found a moderate positive correlation (Spearman’s r = 0.5028, p< 0.05) between pigment grades assessed by these two modalities. These findings suggest that dermoscopic pigmentation reflects both melanin distribution and tumour architecture at the histopathological level. These quantitative correlations advance prior qualitative descriptions of structural associations.

The moderate association may stem from the fact that dermoscopic colour reflects not only melanin density but also its depth and the optical properties of surrounding tissue.^23,24 For instance, leaf-like areas and spoke wheel structures, frequent in superficial BCCs, represent pigmented tumour nests clustered at the dermo-epidermal junction (DEJ) or within the superficial papillary dermis. Similarly, large blue-grey ovoid nests represent melanised tumour islands deeper in the dermis. The correlation between dermoscopic features and histopathology of BCC has been summarised in Table 3.^25-28

This likely explains the higher detection rate of pigmentation by dermoscopy (96.7%) compared to histopathology (66.7%), particularly in infiltrative BCCs, particularly in subtypes like infiltrative BCC (dermoscopic: 90.0%, histopathological: 40.0%), where histopathology may underestimate melanisation due to sampling limitations. This discrepancy between the higher pigmentation detection rate by dermoscopy (96.7%) compared to histopathology (66.7%), particularly pronounced in infiltrative BCC (90.0% vs. 40.0%), is likely attributable to the fundamental differences in the techniques. Dermoscopic optical sensitivity allows it to detect minute melanin deposition within the epidermis or superficial dermis that may fall below the threshold of detection or be missed in routine histopathological evaluation.^12,29 Limitations inherent to histopathology include the sensitivity of standard staining methods (e.g., haematoxylin-eosin), potential sectioning artifacts obscuring focal pigmentation, and melanin distribution that might be diffuse or sparse.²⁸ Furthermore, dermoscopy can visualise pigmented structures such as melanin-laden macrophages in situ, which might not be specifically commented upon or categorised as ‘tumour pigmentation in the histopathological report, focusing on tumour nests themselves.³⁰ Crucially, dermoscopy detects subclinical pigmentation invisible to the naked eye, underscoring its role in the early assessment of BCC.¹⁰

Unlike most prior studies that focused on structural correlations (e.g., the association between blue-grey nests and nodular BCC), this study first quantifies the pigment content correlation between dermoscopy and histopathology. The observed moderate correlation (r=0.5028) advanced the field from qualitative descriptions to quantitative analysis, providing a new metric for non-invasive assessment of BCC. This quantitative approach may help differentiate BCC from melanoma, as previous studies have shown that dermoscopic pigment patterns in BCC (e.g., spoke-wheel-like structures) are distinct from melanoma-specific patterns such as radial streaming.

The poor interobserver agreement (Cohen’s kappa = −0.032, p > 0.05) highlighted the subjectivity of visual pigment grading. This discrepancy likely arose from the lack of objectivity in pigment assessment and the quantitative differences in pigment distribution across cases. To address this, future studies should develop standardised grading protocols (e.g., using image analysis software to measure pigment area).

Consistent with prior reports, leaf-like areas and spoke-wheel-like structures were significantly more frequent on the trunk and limbs (Area M+L, p<0.05),²² while arborising vessels, ulceration, and blue-white veil were more prevalent in the “mask area” (Area H, p<0.05).³¹

Dotted vessels and leaf-like areas were associated with a shorter disease course (<24 months, p<0.05), aligning with findings linking dotted vessels to early-stage and superficial BCCs.³² The characteristics like large blue-grey ovoid nests, multiple blue-grey dots and globules, dotted vessels, and blue-white veil showed statistically significant variations between males and females (p<0.05). The higher prevalence of BCC in women (56.7%) compared to men (43.3%) may be attributable to women’s greater attention to appearance, leading to earlier detection.²² The gender-specific distribution of blue-grey ovoid nests (male predominance, OR=3.490, 95% CI: 1.2–10.1, p = 0.010) may be linked to hormonal influences on dermal melanocyte activity or anatomical differences (e.g., thicker skin in males enhancing deep pigment detection).

While prior studies have reported associations between dermoscopic structures (e.g., ulceration, arborising vessels) and high-risk BCCs, this study uniquely quantifies pigment content rather than solely focusing on structural features. Our quantitative approach, combined with established knowledge of subtype-specific pigmentation patterns, suggests that pigment quantification may hold potential for improving risk stratification, such as distinguishing heavily pigmented nodular BCCs from less pigmented infiltrative subtypes. This finding supports dermoscopy as a bridge between clinical examination and histopathology, enabling non-invasive monitoring of tumour progression.

Limitations

The limitations of our study include a small sample size and geographic homogeneity (predominantly Asian individuals with Fitzpatrick skin types III-IV), which may limit generalisability to other populations.

Conclusion

In conclusion, while dermoscopy is a valuable tool for diagnosing BCC and quantitatively assessing its pigmentation (reflecting underlying melanin and tumour aggregation), our findings confirm that dermoscopic assessment has limitations and does not entirely replace histopathology. Therefore, the diagnosis and management of BCC should be based on a comprehensive approach integrating dermoscopy, clinical examination, and histopathological evaluation when indicated.