Non-invasive Imaging Modalities for Breslow Thickness Estimation in Melanoma: A Scoping Review.

BACKGROUND

Breslow thickness (BT) is the most important prognostic biomarker in cutaneous melanoma and directly influences surgical management, sentinel lymph node biopsy decisions and survival outcomes. Several non-invasive imaging modalities have been investigated for preoperative BT estimation; however, their relative performance and clinical applicability remain unclear.

OBJECTIVE

To evaluate the agreement between non-invasive imaging-derived and histopathologically measured BT and to compare the technical characteristics, diagnostic performance, and clinical utility of currently available imaging modalities for preoperative melanoma staging.

METHODS

A scoping review was conducted according to Joanna Briggs Institute methodology and reported following PRISMA-ScR guidelines. MEDLINE, Scopus, and Web of Science were searched from inception to December 1, 2025. Original studies evaluating non-invasive imaging techniques for BT estimation in cutaneous melanoma were included. Imaging modalities assessed comprised dermoscopy, reflectance confocal microscopy (RCM), optical coherence tomography (OCT), ultrasound (US), multispectral imaging (MSI), photoacoustic techniques, magnetic resonance imaging (MRI), single-photon emission computed tomography (SPECT), and artificial intelligence (AI)-based approaches.

RESULTS

Ninety-two studies met the inclusion criteria. Dermoscopy provided indirect assessment of invasiveness, while RCM and OCT offered complementary microstructural information but limited direct thickness estimation. US, particularly high-frequency (≥20 MHz) and ultra-high-frequency (≥50 MHz) systems, showed the strongest and most consistent evidence for direct BT assessment, with correlation coefficients often exceeding 0.90 and high performance across clinically relevant thresholds. Nevertheless, performance varied across devices, probe frequencies, acquisition protocols, and operators, with frequent overestimation and reduced accuracy in selected subgroups (in situ and thin melanomas). MSI showed moderate-to-good classification performance, while photoacoustic techniques yielded promising results but limited to small cohorts. Evidence for MRI and SPECT was scarce. AI-based approaches achieved moderate-to-good predictive performance but did not consistently outperform experienced clinicians, with generalizability limited by dataset and methodological heterogeneity.

CONCLUSIONS

US currently represents the most extensively validated and clinically applicable non-invasive modality for preoperative BT estimation. A hierarchical imaging approach appears most appropriate, with dermoscopy as the first-line diagnostic tool, RCM and OCT providing additional lesion characterization, and HFUS enabling complementary depth assessment in lesions strongly suspicious for melanoma. Emerging photoacoustic and AI-based approaches require further validation, while histopathology remains the reference standard for definitive diagnosis and BT measurement.