Abstract
Background: Multimodal noninvasive imaging is a novel technique in dermatology. Its purpose is to overcome the limitations of unimodal techniques. This is done by combining higher resolution images with deeper imaging depth and/or tissue specificity within a single device for improvement of diagnosis and management of skin lesions.
Objective: Our aim was to systematically review multimodal imaging devices currently used in dermatology clinics and their diagnostic accuracy of skin lesions.
Methods: A comprehensive search was conducted in October 2022 for studies on multimodal imaging technologies in vivo, in human subjects, used in dermatology. An additional search was made in March 2024. Four databases were selected: MEDLINE, Embase, CENTRAL, and Web of Science. This review was registered with the international Prospective Register of Systematic Reviews (PROSPERO record number CRD42022364864). The search strategy used Medical Subject Headings (MeSH) and keywords from two concepts: Multimodal Imaging and Dermatological Conditions. Reference lists were also searched for relevant studies. Selected studies included multimodal techniques in neoplastic, inflammatory, and pigmentary skin disorders. Ex vivo imaging, non-human studies, and non-clinical settings were excluded. Data extraction and quality assessment were performed using QUADAS and STARD criteria. Data extraction was conducted by one researcher and then reviewed by an additional researcher. A third researcher resolved any disagreements. The statistical analysis involved diagnostic accuracy meta-analysis, subgroup comparisons (multimodal vs single modal, multimodal vs multimodal), and SROC curves. The primary outcomes were the diagnostic accuracy, sensitivity, and specificity of multimodal imaging devices in diagnosing dermatological conditions.
Results: The analysis included 92 studies, predominantly case reports or series (83.7%), with basal cell carcinoma (BCC) being the most frequently imaged lesion (11.8%). The studies had a moderate risk of bias (QUADAS score: 0.6) with histology as the reference test in most cases. Meta-analysis showed multimodal devices have high sensitivity and specificity for BCC diagnosis. These results indicate reliable diagnostic accuracy.
Conclusions: In a clinical setting, multimodal imaging can be useful to perform bedside diagnosis and management of skin lesions.
Plain language summary
This systematic review and meta-analysis systematically evaluated published data on multimodal imaging techniques and assessed their role in diagnosing of skin lesions in a clinical setting. We also performed a meta-analysis to evaluate the diagnostic accuracy of each device for various skin lesions and assessed their risk of bias. Skin lesions, including cancerous and noncancerous lesions, are common; therefore, early and accurate diagnosis is crucial for effective treatment. The current standard of care is biopsy, which is invasive and painful. Noninvasive imaging offers a painless, bedside diagnostic solution for skin examination, reducing unnecessary biopsies and aiding in treatment management. However, most existing techniques rely on a single technology and have limitations. Multimodal imaging, which combines two or more noninvasive imaging techniques, addresses these limitations. We found that basal cell carcinoma was the most commonly studied lesion, and the most widely used device was the line-field confocal optical coherence tomography (LC-OCT). The studies included in this review generally had a moderate risk of bias. A limitation of this review is the scarcity of studies performed on large cohorts, with only a few studies assessing diagnostic accuracy. Nonetheless, we conclude that multimodal imaging is the future of noninvasive diagnostics and has the potential to significantly improve patient care.