Veronica Rotemberg's research while affiliated with Memorial Sloan Kettering Cancer Center and other places

... As RCM images are digital in nature, they can be read remotely. Remote interpretation can be achieved via two methods: a standard store-and-forward (SAF) method [108,109] and a new live interactive method (LIM) tele-RCM [110]. With the SAF method, images are transferred to a remote expert reader after they are acquired, while with the LIM tele-RCM, the expert joins the imaging session with access to the screen in real-time. ...

... 60-62 A recent study by Daneshjou et al. pointed out the challenge of developing an unbiased and accurate data set for AI training, and the importance of fine-tuning AI models to close the performance gap between light and dark skin tones. 60 Future research and development should emphasize the importance of training AI software to recognize and accurately diagnose dermatoscopic images in patients of all skin tones. ...

... It may be assumed that the incorporation of these resources could offer clinical support to health professionals to speed up decision-making regarding diagnosis and management of different diseases [1,11,[16][17][18]. The combination of interactive documentation with metadata, image annotations with text, tables, graphics, and hyperlinks optimize communication between medical professionals [2,19]. The incorporation of these resources with artificial intelligence has been gaining prominence in the health area [2]. ...

... In most cases, the anatomical context of such images is lost due to the exclusion of surrounding structures, while the primary focus of the image is the lesion. Furthermore, with the rapid adoption of digital skin imaging applications, the utilization of smartphoneacquired images in dermatology have also increased proportionally [123]. While many studies have proposed methods to detect melanomas from inconsistent dermoscopy images, most of them produce localized results that cannot be used universally due to the acquisitive conditions they are trained on, such as isolated datasets and specific illumination conditions, etc. [124]. ...

... Skin cancer is one of the most common global public health problems, with increased mortality rates and morbidity and treatment costs annually (Thuncharoen et al., 2013;Yahya et al., 2019;Hwang et al., 2020). Despite the abundance of data on the presentation of skin cancer in white individuals, there is a paucity of data regarding disease morphology and risk factors in darker-skinned individuals because the incidence of skin cancer is relatively higher in white individuals (Gordon et al., 2022;Manci et al., 2022). The leading cause of skin cancer is UV radiation. ...

... The algorithms performed better than experts in most categories, with the exception of actinic keratoses (similar accuracy on average) and images from categories not included in algorithmic training data. 53 Stiff et al. evaluated the advantages and challenges of AI in the detection of melanoma using dermoscopy in a 2022 review article. They concluded that AI may offer benefits beyond diagnosis; it can detect features that predict melanoma prognosis such as likelihood of response to immune checkpoint inhibitors and may be able to classify patients as 'high risk' (which coincides with a significantly decreased chance for progression-free survival). ...

... Trends in melanoma incidence need to be interpreted in light of changing surveillance practices. In the United States [55,56], Australia [57], and Europe [58,59] there has been a much greater increase in the incidence of in situ (confined to the epidermis) and thin melanomas compared with thick melanomas. The increase in melanoma incidence has also greatly outstripped increases in the mortality rate. ...

... Quality was assessed using the Checklist for Evaluation of Image-Based Artificial Intelligence Reports in Dermatology (CLEAR Derm Consensus Guidelines) [22]. This 25-point checklist offers comprehensive recommendations on factors critical to the development, performance and application of image-based AI algorithms in dermatology [22]. ...

... LC-OCT is gaining considerable interest in (dermato-) oncology, as the results achieved by combining RCM and OCT are encouraging by compensating the limitations of each device. Of particular importance, there are some reports highlighting that LC-OCT may guide the detection of the most relevant diagnostic or prognostic areas in vivo, allowing for accurate and targeted biopsies and hence precise downstream histopathology and molecular profiling in cancers [111]. OCT and RCM-assisted sampling might also play a pivotal role in monitoring therapeutic responses at a cellular level by tracking the tumor mutational burden or evaluating the expression of immune biomarkers such as PD-L1 in cutaneous malignancies [111]. ...

... Some of these instruments allow a faster and more accurate diagnosis of skin neoplasms, which is necessary to ensure adequate treatment of the patient. At the same time, they require specific, lengthy training and may increase costs to health systems if used inappropriately [131][132][133]. ...