An RNA vaccine drives immunity in checkpoint-inhibitor-treated melanoma.

Treating patients who have cancer with vaccines that stimulate a targeted immune response is conceptually appealing, but cancer vaccine trials have not been successful in late-stage patients with treatment-refractory tumours1,2. We are testing melanoma FixVac (BNT111)-an intravenously administered liposomal RNA (RNA-LPX) vaccine, which targets four non-mutated, tumour-associated antigens that are prevalent in melanoma-in an ongoing, first-in-human, dose-escalation phase I trial in patients with advanced melanoma (Lipo-MERIT trial, ClinicalTrials.gov identifier NCT02410733). We report here data from an exploratory interim analysis that show that melanoma FixVac, alone or in combination with blockade of the checkpoint inhibitor PD1, mediates durable objective responses in checkpoint-inhibitor (CPI)-experienced patients with unresectable melanoma. Clinical responses are accompanied by the induction of strong CD4+ and CD8+ T cell immunity against the vaccine antigens. The antigen-specific cytotoxic T-cell responses in some responders reach magnitudes typically reported for adoptive T-cell therapy, and are durable. Our findings indicate that RNA-LPX vaccination is a potent immunotherapy in patients with CPI-experienced melanoma, and suggest the general utility of non-mutant shared tumour antigens as targets for cancer vaccination.

Dimethyl fumarate treatment in relapsed and refractory cutaneous T-cell lymphoma: a multicenter phase 2 study.

Targeted therapies for cutaneous T-cell lymphoma (CTCL) are limited and curative approaches are lacking. Furthermore, relapses and drug induced side effects are major challenges in the therapeutic management of patients with CTCL, creating an urgent need for new and effective therapies. Pathologic constitutive NF-κB activity leads to apoptosis resistance in CTCL cells and, thus, represents a promising therapeutic target in CTCL. In a preclinical study we showed the potential of dimethyl fumarate (DMF) to block NF-κB and, specifically, kill CTCL cells. To translate these findings to applications in a clinical setting, we performed a multicentric phase 2 study evaluating oral DMF therapy in 25 patients with CTCL stages Ib to IV over 24 weeks (EudraCT number 2014-000924-11/NCT number NCT02546440). End points were safety and efficacy. We evaluated skin involvement (using a modified severity weighted assessment tool [mSWAT]), pruritus, quality of life, and blood involvement, if applicable, as well as translational data. Upon skin analysis, 7 of 23 (30.4%) patients showed a response with >50% reduction in the mSWAT score. Patients with high tumor burden in the skin and blood responded best to DMF therapy. Although not generally significant, DMF also improved pruritus in several patients. Response in the blood was mixed, but we confirmed the NF-κB-inhibiting mechanism of DMF in the blood. The overall tolerability of the DMF therapy was very favorable, with mostly mild side effects. In conclusion, our study presents DMF as an effective and excellently tolerable therapeutic option in CTCL to be further evaluated in a phase 3 study or real-life patient care as well as in combination therapies. This trial was registered at www.clinicaltrials.gov as #NCT02546440.

Influence of adjuvant therapies on organ-specific recurrence of cutaneous melanoma: A multicenter study on 1383 patients of the prospective DeCOG registry ADOReg.

This study investigated whether adjuvant treatments in stage III cutaneous melanoma (CM) influenced patterns of recurrence. Patients with primary (n = 1033) or relapsed CM (n = 350) who received adjuvant therapies with Nivolumab (N), Pembrolizumab (P), or Dabrafenib and Trametinib (D + T) were extracted from the prospective multicenter real-world skin cancer registry ADOReg. Endpoints were progression-free survival (PFS), distant metastasis-free survival (DMFS), organ-specific DMFS, and overall survival (OS). For primary cases, D + T indicated an improved PFS (1- and 2-year PFS: 90.9%; 82.7%) as compared to P (81.0%, 73.9%; p = .0208), or N (83.8%, 75.2%; p = .0539). BRAF-mutated(mut) CM demonstrated significantly lower PFS (p = .0022) and decreased DMFS (p = .0580) when treated with immune checkpoint inhibitor (ICI) instead of D + T. Besides, NRAS-mut CM tended to perform worse than wt CM upon ICI (PFS: p = .1349; DMFS: p = .0540). OS was similar between the groups. Relapsed cases showed decreased PFS, DMFS, and OS in comparison to primary (all: p < .001), without significant differences between the subgroups. Organ-specific DMFS was significantly altered for primary cases with bone (p = .0367) or brain metastases (p = .0202). In relapsed CM, the frequency of liver (D + T: 1.5%; P: 12%; N: 9%) and LN metastases (D + T: 1.5%; P: 12%; N: 10.2%) was significantly lower with adjuvant D + T than ICI. NRAS-mut CM showed increased recurrence in primary and relapsed cases. These data show that adjuvant D + T is superior to ICI in primary BRAF-mut CM.

Generation of Myeloid-Derived Suppressor Cells Mediated by MicroRNA-125a-5p in Melanoma.

The ability of tumor-derived extracellular vesicles (EVs) to modulate the function of myeloid cells is widely recognized. Hence, a comprehensive understanding of the distinct components associated with EVs and the signals that they deliver to myeloid cells could provide potential approaches to impede the immunosuppression by myeloid-derived suppressor cells (MDSCs). We investigated melanoma EV-associated microRNAs (miRs) using the RET transgenic melanoma mouse model and simulated their transfer to normal myeloid cells by transfecting immature mouse myeloid cells and human monocytes. We observed elevated levels of miR-125a-5p, -125b-5p, and let-7e-5p in mouse melanoma-infiltrating MDSCs. In addition, miR-125a-5p levels in the tumor microenvironment correlated with mouse melanoma progression. The delivery of miR-125a-5p, alone or in combination with let-7e-5p and miR-99b-5p from the same genomic cluster, to normal myeloid cells resulted in their conversion to MDSC-like cells. Our findings indicate that miR-125a-5p could modulate myeloid cell activation in the melanoma microenvironment via a NF-κB-dependent mechanism.

S1-Guideline Sebaceous Carcinoma.

Sebaceous gland carcinomas are rare malignant cutaneous adnexal tumors with sebocytic differentiation. The typical predilection area is the head and neck region, where sebaceous gland carcinomas are the most common malignant adnexal tumors of the skin. According to their localization a distinction is made between periocular and extraocular sebaceous gland carcinomas. Muir-Torre syndrome (MTS) should always be ruled out if it is suspected. In terms of prognosis, sebaceous gland carcinomas are potentially aggressive tumors with a clear tendency to recur and metastasize. Only small extraocular sebaceous gland carcinomas that have been completely resected have a very good prognosis. Sebaceous gland carcinomas most frequently metastasize lymphogenously to regional or distant lymph nodes; organ metastasis occurs less frequently. Periocular sebaceous gland carcinomas have a higher metastasis rate (up to 15%) than extraocular sebaceous gland carcinomas (up to 2%). Complete micrographically controlled surgery (MCS) of the primary tumor is the therapy of first choice, regardless of periocular or extraocular localization. Adjuvant or therapeutic radiotherapy may be considered. There is currently no established standard therapy for advanced, inoperable, or metastatic sebaceous gland carcinomas. Local procedures and systemic therapies such as chemotherapy or immunotherapy can be considered. The procedure should be determined individually by an interdisciplinary tumor board. Close follow-up care is recommended for these potentially aggressive carcinomas.

A self-supervised vision transformer to predict survival from histopathology in renal cell carcinoma.

PURPOSE: To develop and validate an interpretable deep learning model to predict overall and disease-specific survival (OS/DSS) in clear cell renal cell carcinoma (ccRCC). METHODS: Digitised haematoxylin and eosin-stained slides from The Cancer Genome Atlas were used as a training set for a vision transformer (ViT) to extract image features with a self-supervised model called DINO (self-distillation with no labels). Extracted features were used in Cox regression models to prognosticate OS and DSS. Kaplan-Meier for univariable evaluation and Cox regression analyses for multivariable evaluation of the DINO-ViT risk groups were performed for prediction of OS and DSS. For validation, a cohort from a tertiary care centre was used. RESULTS: A significant risk stratification was achieved in univariable analysis for OS and DSS in the training (n = 443, log rank test, p < 0.01) and validation set (n = 266, p < 0.01). In multivariable analysis, including age, metastatic status, tumour size and grading, the DINO-ViT risk stratification was a significant predictor for OS (hazard ratio [HR] 3.03; 95%-confidence interval [95%-CI] 2.11-4.35; p < 0.01) and DSS (HR 4.90; 95%-CI 2.78-8.64; p < 0.01) in the training set but only for DSS in the validation set (HR 2.31; 95%-CI 1.15-4.65; p = 0.02). DINO-ViT visualisation showed that features were mainly extracted from nuclei, cytoplasm, and peritumoural stroma, demonstrating good interpretability. CONCLUSION: The DINO-ViT can identify high-risk patients using histological images of ccRCC. This model might improve individual risk-adapted renal cancer therapy in the future.

Personalized RNA mutanome vaccines mobilize poly-specific therapeutic immunity against cancer.

T cells directed against mutant neo-epitopes drive cancer immunity. However, spontaneous immune recognition of mutations is inefficient. We recently introduced the concept of individualized mutanome vaccines and implemented an RNA-based poly-neo-epitope approach to mobilize immunity against a spectrum of cancer mutations. Here we report the first-in-human application of this concept in melanoma. We set up a process comprising comprehensive identification of individual mutations, computational prediction of neo-epitopes, and design and manufacturing of a vaccine unique for each patient. All patients developed T cell responses against multiple vaccine neo-epitopes at up to high single-digit percentages. Vaccine-induced T cell infiltration and neo-epitope-specific killing of autologous tumour cells were shown in post-vaccination resected metastases from two patients. The cumulative rate of metastatic events was highly significantly reduced after the start of vaccination, resulting in a sustained progression-free survival. Two of the five patients with metastatic disease experienced vaccine-related objective responses. One of these patients had a late relapse owing to outgrowth of ß2-microglobulin-deficient melanoma cells as an acquired resistance mechanism. A third patient developed a complete response to vaccination in combination with PD-1 blockade therapy. Our study demonstrates that individual mutations can be exploited, thereby opening a path to personalized immunotherapy for patients with cancer.

Flagellin-Induced Immune Response in Human-Induced Pluripotent Stem Cell-Derived Cardiomyocytes.

Pathogen-associated molecular patterns (PAMPs) are involved in the pathogenesis of septic cardiomyopathy through a toll-like receptor (TLR)-mediated immune response. Human-induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) can reflect the innate immune abilities of cardiomyocytes. Therefore, hiPSC-CMs may provide an attractive tool with which to study PAMP-induced alterations in cardiomyocytes. HiPSC-CMs from two different healthy donors were exposed to the PAMP flagellin (FLA) at different doses and exposure times. Alterations in the expression levels of distinct inflammation-associated cytokines, intracellular inflammation pathways including TLR5 downstream signaling, reactive oxygen species levels and surface antigen composition were assessed using PCR, ELISA and FACS techniques. Higher doses of flagellin increased the expression levels of inflammation-associated cytokines like TNFα (p < 0.01) and downstream signaling molecules like caspase-8 (p < 0.05). TLR5 expression (p < 0.01) and TLR5 fluorescence proportion (p < 0.05) increased in hiPSC-CMs after prolonged FLA exposure. FLA-induced innate immune response processes in cardiomyocytes might be detectable with an hiPSC-CMs-based in vitro model.

Nuclear AREG affects a low-proliferative phenotype and contributes to drug resistance of melanoma.

AMPHIREGULIN (AREG) is a multifaceted molecule, which acts not only as an extracellular ligand for EGF receptor (EGFR), but also as an intracellular signaling molecule. It remains elusive, however, whether AREG has a tumor suppressive or oncogenic role in melanoma. Here, we found that several melanoma cell lines express AREG, but the expression does not correlate with that of EGFR. Recombinant AREG and the neutralizing antibody experiments showed that intracellular AREG plays an important role in melanoma, implying a divergent function of AREG in addition to the role as a ligand for EGFR. Further investigation of this mechanism revealed that particularly nuclear-localized AREG regulates IGF-1R, P21 (Cip1/Waf1), TP53 and JARID1B protein accumulation in the nucleus. Furthermore, manipulation of nuclear AREG levels has influence on heterochromatin condensation (HP1beta, SETDB1) and trimethylation of histones H3K9 and H3K4. As these molecules correspond to previously identified markers for slow-cycling drug resistant cells, we speculate that nuclear AREG predisposes cells to resistance to therapy. According to the hypothesis, we detected the accumulation of AREG in the nucleus of SK-Mel-28-VR, which was cultured under Vemurafenib (VR) selection pressure, and this correlates with JARID1B expression. Here, knockdown of AREG makes the previously resistant cells more sensitive to VR treatment, resulting in inhibited proliferation. Taken together, we suggest that nuclear AREG affects a slow-cycling phenotype and increases resistance to VR, raising a possibility that AREG might be a potential therapeutic target for resistance in melanoma.

Clinical determinants of long-term survival in metastatic uveal melanoma.

This study aimed to identify prognostic factors in patients with metastatic uveal melanoma (UM) that were associated with long-term survival in a real-world setting. A total of 94 patients with metastatic UM were included from German skin cancer centers and the German national skin cancer registry (ADOReg). Data were analyzed for the response to treatment, progression-free survival, and overall survival (OS). Prognostic factors were explored with univariate Cox regression, log-rank, and χ2-tests. Identified factors were subsequently validated after the population was divided into two cohorts of short-term survival (< 2 years OS, cohort A, n = 50) and long-term survival (> 2 years OS, cohort B, n = 44). A poor ECOG performance status (hazard ratio [HR] 2.0, 95% confidence interval [CI] 1.0-3.9) and elevated serum LDH (HR 2.0, 95% CI 1.0-3.8) were associated with a poor OS, whereas a good response to immune checkpoint blockade (ICB, p < 0.001), radiation therapy (p < 0.001), or liver-directed treatments (p = 0.01) were associated with a prolonged OS. Long-term survivors (cohort B) showed a higher median number of organs affected by metastasis (p < 0.001), while patients with liver metastases only were more common in cohort A (40% vs. 9%; p = 0.002). A partial response to ICB was observed in 16% (12/73), being 21% (8/38) for combined ICB, 17% (1/6) for single CTLA4 inhibition, and 10% (3/29) for single PD1 inhibition. One complete response occurred in cohort B with combined ICB. We conclude that the response to ICB and the presence of extrahepatic disease were favorable prognostic factors for long-term survival.

Effectiveness, Safety and Utilization of Vismodegib for Locally Advanced Basal Cell Carcinoma Under Real-world Conditions: Non-interventional Cohort Study JONAS.

Most patients with advanced basal cell carcinomas (BCCs) may not benefit sufficiently from standard treatment comprising surgery and radiation. Vismodegib, an oral selective hedgehog pathway inhibitor, is approved for treatment of patients with locally advanced BCC inappropriate for surgery or radiotherapy, or for patients with symptomatic metastatic BCC. In order to enhance understanding of the effectiveness, safety and utilization of vismodegib in clinical practice in Germany, a non-interventional study, JONAS, was conducted. A total of 53 patients with locally advanced BCC who initiated treatment with vismodegib between 2016 and 2018 were included in the study, which was embedded in the German ADOReg skin cancer registry. Duration of response, the primary endpoint, was 12.4 months, progression-free survival 32.2 months and overall response rate 77.4%. Most adverse events were mild to moderate. Overall, results confirmed previous findings, demonstrating favourable responses and manageable safety of vismodegib in patients with locally advanced BCC in clinical practice.

Melanoma Cellular Plasticity.

Despite the advances of modern medicine and the development of innovative and promising new therapeutic strategies for the treatment of the numerous types of cancer, far too many patients still lose the battle against these devastating diseases [...].

S1-guideline atypical fibroxanthoma (AFX) and pleomorphic dermal sarcoma (PDS).

Atypical fibroxanthoma (AFX) and pleomorphic dermal sarcoma (PDS) are rare cutaneous neoplasms representing histomorphological, genetic as well as epigenetic variants of a disease spectrum. Both tumors typically manifest as nonspecific, often ulcerated, skin- to flesh-colored nodules in chronically sun-damaged skin of elderly male patients. AFX is a rather well demarcated, often rapidly growing tumor. PDS tumors are poorly circumscribed and are characterized by aggressive infiltrative growth. Fast as well as slow growth behavior has been reported for both tumors. Histologically, both are composed of spindle-shaped and epithelioid tumor cells with pleomorphic nuclei as well as atypical multinucleated giant cells. Atypical mitoses are common. In contrast to AFX, PDS involves relevant parts of the subcutis and shows areas of tumor necrosis and/or perineural infiltration. Due to the poorly differentiated nature of AFX/PDS (Grade 3), histopathologically similar cutaneous sarcomas, undifferentiated carcinomas, melanomas and other diseases have to be excluded by immunohistochemical analysis. The treatment of choice is micrographically controlled surgery. In cases of AFX, a cure can be assumed after complete excision. Local recurrence rates are low as long as PDS tumors are surgically removed with a safety margin of 2 cm. Metastasis is rare and mostly associated with very thick or incompletely excised tumors; it mainly affects the skin and lymph nodes. Distant metastasis is even more rare. No approved and effective systemic therapy has been established.

SOX2 in development and cancer biology.

The transcription factor SOX2 is essential for embryonic development and plays a crucial role in maintaining the stemness of embryonic cells and various adult stem cell populations. On the other hand, dysregulation of SOX2 expression is associated with a multitude of cancer types and it has been shown that SOX2 positively affects cancer cell traits such as the capacity to proliferate, migrate, invade and metastasize. Moreover, there is growing evidence that SOX2 mediates resistance towards established cancer therapies and that it is expressed in cancer stem cells. These findings indicate that studying the role of SOX2 in the context of cancer progression could lead to the development of new therapeutic options. In this review, the current knowledge about the role of SOX2 in development, maintenance of stemness, cancer progression and the resistance towards cancer therapies is summarized.

Tumor promoting capacity of polymorphonuclear myeloid-derived suppressor cells and their neutralization.

Myeloid-derived suppressor cells (MDSC) represent a highly immunosuppressive population that expands in tumor bearing hosts and inhibits both T and NK cell antitumor effector functions. Among MDSC subpopulations, the polymorphonuclear (PMN) one is gaining increasing interest since it is a predominant MDSC subset in most cancer entities and inherits unique properties to facilitate metastatic spread. In addition, further improvement in distinguishing PMN-MDSC from neutrophils has contributed to the design of novel therapeutic approaches. In this review, we summarize the current view on the origin of PMN-MDSC and their relation to classical neutrophils. Furthermore, we outline the metastasis promoting features of these cells and promising strategies of their targeting to improve the efficacy of cancer immunotherapy.

FOXD1 promotes dedifferentiation and targeted therapy resistance in melanoma by regulating the expression of connective tissue growth factor.

Metastatic melanoma is an aggressive skin cancer and associated with a poor prognosis. In clinical terms, targeted therapy is one of the most important treatments for patients with BRAFV600E -mutated advanced melanoma. However, the development of resistance to this treatment compromises its therapeutic success. We previously demonstrated that forkhead box D1 (FOXD1) regulates melanoma migration and invasion. Here, we found that FOXD1 was highly expressed in melanoma cells and was associated with a poor survival of patients with metastatic melanoma. Upregulation of FOXD1 expression enhanced melanoma cells' resistance to vemurafenib (BRAF inhibitor [BRAFi]) or vemurafenib and cobimetinib (MEK inhibitor) combination treatment whereas loss of FOXD1 increased the sensitivity to treatment. By comparing gene expression levels between FOXD1 knockdown (KD) and overexpressing (OE) cells, we identified the connective tissue growth factor (CTGF) as a downstream factor of FOXD1. Chromatin immunoprecipitation and luciferase assay demonstrated the direct binding of FOXD1 to the CTGF promoter. Similar to FOXD1, knockdown of CTGF increased the sensitivity of BRAFi-resistant cells to vemurafenib. FOXD1 KD cells treated with recombinant CTGF protein were less sensitive towards vemurafenib compared to untreated FOXD1 KD cells. Based on these findings, we conclude that FOXD1 might be a promising new diagnostic marker and a therapeutic target for the treatment of targeted therapy resistant melanoma.

IL-6 as a major regulator of MDSC activity and possible target for cancer immunotherapy.

Myeloid-derived suppressor cells (MDSC) are generated during tumor progression and suppress the anti-tumor functions of T and natural killer (NK) cells. Their enrichment is associated with a bad prognosis and a worse outcome of immunotherapy in cancer patients. The cytokine interleukin (IL)-6 was found to be a crucial regulator of MDSC accumulation and activation as well as a factor, stimulating tumor cell proliferation, survival, invasiveness and metastasis. Accordingly, IL-6 can serve as a negative prognostic marker in cancer. On the other hand, this cytokine is also involved in T cell activation. This review discusses the pleiotropic effects of IL-6 on immune cell populations that are critical for tumor development, such as MDSC and T cells, and summarizes the data on targeting IL-6 or IL-6 receptor (IL-6R) for tumor immunotherapy to block MDSC-mediated immunosuppression in cancer patients.

Potential therapeutic effect of low-dose paclitaxel in melanoma patients resistant to immune checkpoint blockade: A pilot study.

The low dose application of chemotherapeutic agents such as paclitaxel was previously shown to initiate anti-tumor activity by neutralizing myeloid-derived suppressor cells (MDSCs) in melanoma mouse models. Here, we investigated immunomodulating effects of low-dose paclitaxel in 9 metastatic melanoma patients resistant to prior treatments. Three patients showed response to therapy (two partial responses and one stable disease). In responding patients, paclitaxel decreased the frequency and immunosuppressive pattern of MDSCs in the peripheral blood and skin metastases. Furthermore, paclitaxel modulated levels of inflammatory mediators in the serum. In addition, responders displayed enhanced frequencies of tumor-infiltrating CD8+ T cells and their activity indicated by the upregulation of CD25 and TCR ζ-chain expression. Our study suggests that low-dose paclitaxel treatment could improve clinical outcome of some advanced melanoma patients by enhancing anti-tumor immunity and might be proposed for combined melanoma immunotherapy.

Deep learning approach to predict lymph node metastasis directly from primary tumour histology in prostate cancer.

OBJECTIVE: To develop a new digital biomarker based on the analysis of primary tumour tissue by a convolutional neural network (CNN) to predict lymph node metastasis (LNM) in a cohort matched for already established risk factors. PATIENTS AND METHODS: Haematoxylin and eosin (H&E) stained primary tumour slides from 218 patients (102 N+; 116 N0), matched for Gleason score, tumour size, venous invasion, perineural invasion and age, who underwent radical prostatectomy were selected to train a CNN and evaluate its ability to predict LN status. RESULTS: With 10 models trained with the same data, a mean area under the receiver operating characteristic curve (AUROC) of 0.68 (95% confidence interval [CI] 0.678-0.682) and a mean balanced accuracy of 61.37% (95% CI 60.05-62.69%) was achieved. The mean sensitivity and specificity was 53.09% (95% CI 49.77-56.41%) and 69.65% (95% CI 68.21-71.1%), respectively. These results were confirmed via cross-validation. The probability score for LNM prediction was significantly higher on image sections from N+ samples (mean [SD] N+ probability score 0.58 [0.17] vs 0.47 [0.15] N0 probability score, P = 0.002). In multivariable analysis, the probability score of the CNN (odds ratio [OR] 1.04 per percentage probability, 95% CI 1.02-1.08; P = 0.04) and lymphovascular invasion (OR 11.73, 95% CI 3.96-35.7; P < 0.001) proved to be independent predictors for LNM. CONCLUSION: In our present study, CNN-based image analyses showed promising results as a potential novel low-cost method to extract relevant prognostic information directly from H&E histology to predict the LN status of patients with prostate cancer. Our ubiquitously available technique might contribute to an improved LN status prediction.

Reducing the Impact of Confounding Factors on Skin Cancer Classification via Image Segmentation: Technical Model Study.

BACKGROUND: Studies have shown that artificial intelligence achieves similar or better performance than dermatologists in specific dermoscopic image classification tasks. However, artificial intelligence is susceptible to the influence of confounding factors within images (eg, skin markings), which can lead to false diagnoses of cancerous skin lesions. Image segmentation can remove lesion-adjacent confounding factors but greatly change the image representation. OBJECTIVE: The aim of this study was to compare the performance of 2 image classification workflows where images were either segmented or left unprocessed before the subsequent training and evaluation of a binary skin lesion classifier. METHODS: Separate binary skin lesion classifiers (nevus vs melanoma) were trained and evaluated on segmented and unsegmented dermoscopic images. For a more informative result, separate classifiers were trained on 2 distinct training data sets (human against machine [HAM] and International Skin Imaging Collaboration [ISIC]). Each training run was repeated 5 times. The mean performance of the 5 runs was evaluated on a multi-source test set (n=688) consisting of a holdout and an external component. RESULTS: Our findings showed that when trained on HAM, the segmented classifiers showed a higher overall balanced accuracy (75.6% [SD 1.1%]) than the unsegmented classifiers (66.7% [SD 3.2%]), which was significant in 4 out of 5 runs (P<.001). The overall balanced accuracy was numerically higher for the unsegmented ISIC classifiers (78.3% [SD 1.8%]) than for the segmented ISIC classifiers (77.4% [SD 1.5%]), which was significantly different in 1 out of 5 runs (P=.004). CONCLUSIONS: Image segmentation does not result in overall performance decrease but it causes the beneficial removal of lesion-adjacent confounding factors. Thus, it is a viable option to address the negative impact that confounding factors have on deep learning models in dermatology. However, the segmentation step might introduce new pitfalls, which require further investigations.