Refine
Has Fulltext
- yes (22)
Is part of the Bibliography
- yes (22)
Document Type
- Journal article (22)
Language
- English (22) (remove)
Keywords
- melanoma (9)
- deep learning (4)
- artificial intelligence (3)
- machine learning (3)
- BRAF (2)
- PD-1 (2)
- angiogenesis (2)
- immunohistochemistry (2)
- skin cancer (2)
- B7-H1 Antigen (1)
Institute
- Klinik und Poliklinik für Dermatologie, Venerologie und Allergologie (22)
- Urologische Klinik und Poliklinik (4)
- Medizinische Klinik und Poliklinik II (3)
- Pathologisches Institut (3)
- Theodor-Boveri-Institut für Biowissenschaften (3)
- Center for Computational and Theoretical Biology (2)
- Medizinische Klinik und Poliklinik I (2)
- Comprehensive Cancer Center Mainfranken (1)
- Institut für Pharmazie und Lebensmittelchemie (1)
- Institut für diagnostische und interventionelle Neuroradiologie (ehem. Abteilung für Neuroradiologie) (1)
Sonstige beteiligte Institutionen
Background
A basic requirement for artificial intelligence (AI)–based image analysis systems, which are to be integrated into clinical practice, is a high robustness. Minor changes in how those images are acquired, for example, during routine skin cancer screening, should not change the diagnosis of such assistance systems.
Objective
To quantify to what extent minor image perturbations affect the convolutional neural network (CNN)–mediated skin lesion classification and to evaluate three possible solutions for this problem (additional data augmentation, test-time augmentation, anti-aliasing).
Methods
We trained three commonly used CNN architectures to differentiate between dermoscopic melanoma and nevus images. Subsequently, their performance and susceptibility to minor changes (‘brittleness’) was tested on two distinct test sets with multiple images per lesion. For the first set, image changes, such as rotations or zooms, were generated artificially. The second set contained natural changes that stemmed from multiple photographs taken of the same lesions.
Results
All architectures exhibited brittleness on the artificial and natural test set. The three reviewed methods were able to decrease brittleness to varying degrees while still maintaining performance. The observed improvement was greater for the artificial than for the natural test set, where enhancements were minor.
Conclusions
Minor image changes, relatively inconspicuous for humans, can have an effect on the robustness of CNNs differentiating skin lesions. By the methods tested here, this effect can be reduced, but not fully eliminated. Thus, further research to sustain the performance of AI classifiers is needed to facilitate the translation of such systems into the clinic.
Recent research revealed the superiority of artificial intelligence over dermatologists to diagnose melanoma from images. However, 30–50% of all melanomas and more than half of those in young patients evolve from initially benign lesions. Despite its high relevance for melanoma screening, neither clinicians nor computers are yet able to reliably predict a nevus’ oncologic transformation. The cause of this lies in the static nature of lesion presentation in the current standard of care, both for clinicians and algorithms. The status quo makes it difficult to train algorithms (and clinicians) to precisely assess the likelihood of a benign skin lesion to transform into melanoma. In addition, it inhibits the precision of current algorithms since ‘evolution’ image features may not be part of their decision. The current literature reveals certain types of melanocytic nevi (i.e. ‘spitzoid’ or ‘dysplastic’ nevi) and criteria (i.e. visible vasculature) that, in general, appear to have a higher chance to transform into melanoma. However, owing to the cumulative nature of oncogenic mutations in melanoma, a more fine-grained early morphologic footprint is likely to be detectable by an algorithm. In this perspective article, the concept of melanoma prediction is further explored by the discussion of the evolution of melanoma, the concept for training of such a nevi classifier and the implications of early melanoma prediction for clinical practice. In conclusion, the authors believe that artificial intelligence trained on prospective image data could be transformative for skin cancer diagnostics by (a) predicting melanoma before it occurs (i.e. pre-in situ) and (b) further enhancing the accuracy of current melanoma classifiers. Necessary prospective images for this research are obtained via free mole-monitoring mobile apps.
Background
Melanoma is the most dangerous type of skin cancer but is curable if detected early. Recent publications demonstrated that artificial intelligence is capable in classifying images of benign nevi and melanoma with dermatologist-level precision. However, a statistically significant improvement compared with dermatologist classification has not been reported to date.
Methods
For this comparative study, 4204 biopsy-proven images of melanoma and nevi (1:1) were used for the training of a convolutional neural network (CNN). New techniques of deep learning were integrated. For the experiment, an additional 804 biopsy-proven dermoscopic images of melanoma and nevi (1:1) were randomly presented to dermatologists of nine German university hospitals, who evaluated the quality of each image and stated their recommended treatment (19,296 recommendations in total). Three McNemar's tests comparing the results of the CNN's test runs in terms of sensitivity, specificity and overall correctness were predefined as the main outcomes.
Findings
The respective sensitivity and specificity of lesion classification by the dermatologists were 67.2% (95% confidence interval [CI]: 62.6%–71.7%) and 62.2% (95% CI: 57.6%–66.9%). In comparison, the trained CNN achieved a higher sensitivity of 82.3% (95% CI: 78.3%–85.7%) and a higher specificity of 77.9% (95% CI: 73.8%–81.8%). The three McNemar's tests in 2 × 2 tables all reached a significance level of p < 0.001. This significance level was sustained for both subgroups.
Interpretation
For the first time, automated dermoscopic melanoma image classification was shown to be significantly superior to both junior and board-certified dermatologists (p < 0.001).
Background
Several recent publications have demonstrated the use of convolutional neural networks to classify images of melanoma at par with board-certified dermatologists. However, the non-availability of a public human benchmark restricts the comparability of the performance of these algorithms and thereby the technical progress in this field.
Methods
An electronic questionnaire was sent to dermatologists at 12 German university hospitals. Each questionnaire comprised 100 dermoscopic and 100 clinical images (80 nevi images and 20 biopsy-verified melanoma images, each), all open-source. The questionnaire recorded factors such as the years of experience in dermatology, performed skin checks, age, sex and the rank within the university hospital or the status as resident physician. For each image, the dermatologists were asked to provide a management decision (treat/biopsy lesion or reassure the patient). Main outcome measures were sensitivity, specificity and the receiver operating characteristics (ROC).
Results
Total 157 dermatologists assessed all 100 dermoscopic images with an overall sensitivity of 74.1%, specificity of 60.0% and an ROC of 0.67 (range = 0.538–0.769); 145 dermatologists assessed all 100 clinical images with an overall sensitivity of 89.4%, specificity of 64.4% and an ROC of 0.769 (range = 0.613–0.9). Results between test-sets were significantly different (P < 0.05) confirming the need for a standardised benchmark.
Conclusions
We present the first public melanoma classification benchmark for both non-dermoscopic and dermoscopic images for comparing artificial intelligence algorithms with diagnostic performance of 145 or 157 dermatologists. Melanoma Classification Benchmark should be considered as a reference standard for white-skinned Western populations in the field of binary algorithmic melanoma classification.
(1) Background: C-X-C Motif Chemokine Receptor 4 (CXCR4) and Fibroblast Activation Protein Alpha (FAP) are promising theranostic targets. However, it is unclear whether CXCR4 and FAP positivity mark distinct microenvironments, especially in solid tumors. (2) Methods: Using Random Forest (RF) analysis, we searched for entity-independent mRNA and microRNA signatures related to CXCR4 and FAP overexpression in our pan-cancer cohort from The Cancer Genome Atlas (TCGA) database — representing n = 9242 specimens from 29 tumor entities. CXCR4- and FAP-positive samples were assessed via StringDB cluster analysis, EnrichR, Metascape, and Gene Set Enrichment Analysis (GSEA). Findings were validated via correlation analyses in n = 1541 tumor samples. TIMER2.0 analyzed the association of CXCR4 / FAP expression and infiltration levels of immune-related cells. (3) Results: We identified entity-independent CXCR4 and FAP gene signatures representative for the majority of solid cancers. While CXCR4 positivity marked an immune-related microenvironment, FAP overexpression highlighted an angiogenesis-associated niche. TIMER2.0 analysis confirmed characteristic infiltration levels of CD8+ cells for CXCR4-positive tumors and endothelial cells for FAP-positive tumors. (4) Conclusions: CXCR4- and FAP-directed PET imaging could provide a non-invasive decision aid for entity-agnostic treatment of microenvironment in solid malignancies. Moreover, this machine learning workflow can easily be transferred towards other theranostic targets.
Merkel cell carcinoma is a rare, highly aggressive skin cancer with neuroendocrine differentiation. Immune checkpoint inhibition has significantly improved treatment outcomes in metastatic disease with response rates to programmed cell death protein 1/programmed cell death 1 ligand 1 (PD-1/PD-L1) inhibition of up to 62%. However, primary and secondary resistance to PD-1/PD-L1 inhibition remains a so far unsolved clinical challenge since effective and safe treatment options for these patients are lacking.Fourteen patients with advanced (non-resectable stage III or stage IV, Union international contre le cancer 2017) Merkel cell carcinoma with primary resistance to the PD-L1 inhibitor avelumab receiving subsequent therapy (second or later line) with ipilimumab plus nivolumab (IPI/NIVO) were identified in the prospective multicenter skin cancer registry ADOREG. Five of these 14 patients were reported previously and were included in this analysis with additional follow-up. Overall response rate, progression-free survival (PFS), overall survival (OS) and adverse events were analyzed.All 14 patients received avelumab as first-line treatment. Thereof, 12 patients had shown primary resistance with progressive disease in the first tumor assessment, while two patients had initially experienced a short-lived stabilization (stable disease). Six patients had at least one systemic treatment in between avelumab and IPI/NIVO. In total, 7 patients responded to IPI/NIVO (overall response rate 50%), and response was ongoing in 4 responders at last follow-up. After a median follow-up of 18.85 months, median PFS was 5.07 months (95% CI 2.43—not available (NA)), and median OS was not reached. PFS rates at 12 months and 24 months were 42.9% and 26.8 %, respectively. The OS rate at 36 months was 64.3%. Only 3 (21%) patients did not receive all 4 cycles of IPI/NIVO due to immune-related adverse events.In this multicenter evaluation, we observed high response rates, a durable benefit and promising OS rates after treatment with later-line combined IPI/NIVO. In conclusion, our patient cohort supports our prior findings with an encouraging activity of second-line or later-line IPI/NIVO in patients with anti-PD-L1-refractory Merkel cell carcinoma.
Simple Summary
In melanoma patients treated with dabrafenib and trametinib, dose reductions and treatment discontinuations related to adverse events (AE) occur frequently. However, the associations between patient characteristics, AE, and exposure are unclear. Our prospective study analyzed serum (hydroxy-)dabrafenib and trametinib exposure and investigated its association with toxicity and patient characteristics. Additionally, the feasibility of at-home sampling of capillary blood was assessed, and a model to convert capillary blood concentrations to serum concentrations was developed. (Hydroxy-)dabrafenib or trametinib exposure was not associated with age, sex, body mass index, or AE. Co-medication with P-glycoprotein inducers was associated with lower trough concentrations of trametinib but not (hydroxy-)dabrafenib. The applicability of the self-sampling of capillary blood was demonstrated. Our conversion model was adequate for estimating serum exposure from micro-samples. The monitoring of dabrafenib and trametinib may be useful for dose modification and can be optimized by at-home sampling and our new conversion model.
Abstract
Patients treated with dabrafenib and trametinib for BRAF\(^{V600}\)-mutant melanoma often experience dose reductions and treatment discontinuations. Current knowledge about the associations between patient characteristics, adverse events (AE), and exposure is inconclusive. Our study included 27 patients (including 18 patients for micro-sampling). Dabrafenib and trametinib exposure was prospectively analyzed, and the relevant patient characteristics and AE were reported. Their association with the observed concentrations and Bayesian estimates of the pharmacokinetic (PK) parameters of (hydroxy-)dabrafenib and trametinib were investigated. Further, the feasibility of at-home sampling of capillary blood was assessed. A population pharmacokinetic (popPK) model-informed conversion model was developed to derive serum PK parameters from self-sampled capillary blood. Results showed that (hydroxy-)dabrafenib or trametinib exposure was not associated with age, sex, body mass index, or toxicity. Co-medication with P-glycoprotein inducers was associated with significantly lower trough concentrations of trametinib (p = 0.027) but not (hydroxy-)dabrafenib. Self-sampling of capillary blood was feasible for use in routine care. Our conversion model was adequate for estimating serum PK parameters from micro-samples. Findings do not support a general recommendation for monitoring dabrafenib and trametinib but suggest that monitoring can facilitate making decisions about dosage adjustments. To this end, micro-sampling and the newly developed conversion model may be useful for estimating precise PK parameters.
Background: Eosinophils appear to contribute to the efficacy of immunotherapy and their frequency was suggested as a predictive biomarker. Whether this observation could be transferred to patients treated with targeted therapy remains unknown. Methods: Blood and serum samples of healthy controls and 216 patients with advanced melanoma were prospectively and retrospectively collected. Freshly isolated eosinophils were phenotypically characterized by flow cytometry and co-cultured in vitro with melanoma cells to assess cytotoxicity. Soluble serum markers and peripheral blood counts were used for correlative studies. Results: Eosinophil-mediated cytotoxicity towards melanoma cells, as well as phenotypic characteristics, were similar when comparing healthy donors and patients. However, high relative pre-treatment eosinophil counts were significantly associated with response to MAPKi (p = 0.013). Eosinophil-mediated cytotoxicity towards melanoma cells is dose-dependent and requires proximity of eosinophils and their target in vitro. Treatment with targeted therapy in the presence of eosinophils results in an additive tumoricidal effect. Additionally, melanoma cells affected eosinophil phenotype upon co-culture. Conclusion: High pre-treatment eosinophil counts in advanced melanoma patients were associated with a significantly improved response to MAPKi. Functionally, eosinophils show potent cytotoxicity towards melanoma cells, which can be reinforced by MAPKi. Further studies are needed to unravel the molecular mechanisms of our observations.
Background: Rhabdoid melanoma is a rare variant of malignant melanoma with characteristic cytomorphologic features. Due to the potential loss of conventional melanocytic markers, histopathologic diagnosis is often challenging. We hypothesize that immunostaining for PReferentially expressed Antigen in MElanoma (PRAME) might have the potential to uncover the melanocytic origin of these dedifferentiated tumors. Methods: Four cases of rhabdoid primary melanomas were assessed by immunohistochemistry for expression of PRAME and conventional melanocytic markers. Immunohistochemical expression patterns were analyzed in the rhabdoid primaries and, if available, associated metastases. Results: All four cases of rhabdoid primary melanomas showed a strong nuclear positivity for PRAME, while the expression of conventional melanocytic markers S100, MART-1, SOX-10 and HMB-45 was variable between the analyzed cases. Conclusions: In summary, we report four cases of rhabdoid primary melanoma with high to intermediate expression of PRAME despite the partial and variable loss of other melanocytic markers. Hence, PRAME might facilitate the recognition of this highly aggressive entity to avoid misdiagnosis due to histopathologic pitfalls.