European Journal of Cancer

This study compared machine-learning models for predicting recurrence-free survival (RFS), disease-specific survival (DSS), and overall survival (OS) using clinicopathologic data from 1,621 stage I/II primary cutaneous melanoma patients. Our time-to-event models achieved concordance indices of 0.829 for RFS, 0.812 for DSS, and 0.778 for OS. Tumor thickness and mitotic rate were the most important predictors for RFS. Charlson comorbidity score and insurance type were critical for DSS and OS.

ImportanceEasy and accessible biomarkers to predict response to immune checkpoint inhibition (ICI)-treated melanoma are limited. ObjectiveTo evaluate artificial intelligence (AI) detected tumor-infiltrating lymphocytes (TILs) on pretreatment melanoma metastases as a biomarker for response and survival in ICI-treated patients. DesignMulticenter cohort study including patients with advanced melanoma treated with first-line anti-PD1 {+/-} anti-CTLA4 between 2016 and 2023. Median follow-up was 36.3 months. Setting11 melanoma treatment centers in the Netherlands. Participants1,202 patients with advanced cutaneous melanoma. ExposureAll patients received first-line anti-PD1 {+/-} anti-CTLA4. Main Outcome(s) and Measure(s)The percentage of TILs inside manually annotated tumor area in H&E stained pretreatment metastases was determined using the Hover-NeXt model trained and evaluated on an independent melanoma dataset containing 166,718 pathologist-verified manually annotated cells. The primary outcome was objective response rate (ORR); secondary outcomes were progression-free survival (PFS) and overall survival (OS). Correlation with manual TILs, scored according to the guidelines stated by the immune-oncology working group, was evaluated with Spearman correlation coefficients. Logistic regression and Cox proportional regression were conducted, adjusted for age, sex, disease stage, ICI type, BRAF status, brain metastases, LDH level, and performance status. ResultsMetastatic melanoma specimens were available for 1,202 patients, of whom 423 received combination therapy. Median TIL percentage was 9.9% (range 0.3% - 69.4%). A 10% increase in TILs was associated with increased ORR (adjusted OR 1.40 [95% 1.23-1.59]), PFS (adjusted HR 0.85 [95% CI 0.79 - 0.92]) and OS (adjusted HR 0.83 [95% CI 0.76 - 0.91]. Results were consistent for both patients treated with anti-PD1 monotherapy and combination treatment with anti-PD1 plus anti-CTLA4. When comparing manual TIL scoring with AI-detected TILs, associations with response and survival were consistently stronger for AI-detected TILs. Conclusions and RelevanceIn patients with advanced melanoma, higher levels of AI-detected TILs on pre-treatment H&E slides were independently associated with improved ICI response and survival. Given the accessibility of TIL scoring on routine histology, TILs may serve as a predictive biomarker for ICI outcomes. To facilitate broader validation, the Hover-NeXt architecture and model weights are publicly available. Key pointsQuestion: What is the predictive value of artificial intelligence-detected tumor-infiltrating lymphocytes (TILs) for clinical outcomes in patients with advanced melanoma receiving first-line immune checkpoint inhibition? Findings: In this multicenter cohort of 1202 patients, TILs in pretreatment metastases were quantified using a melanoma-specific publicly available AI model trained on an independent dataset. A 10% increase was associated with response (aOR 1.40 [95% CI 1.23-1.59]), progression-free survival (aHR 0.85 [95% CI 0.79-0.92]), and overall survival (aHR 0.83 [95% CI 0.76-0.91]). Associations were independent of clinical predictors. Meaning: AI-detected TILs in pretreatment melanoma metastases independently correlate with response and survival.

Acral melanoma (AM) is a rare subtype of melanoma associated with poor prognosis. However, inconsistent definitions--based on either anatomic location (e.g., palms, soles, subungual areas) or histopathologic subtype (e.g., acral lentiginous melanoma)--complicate prognostication, hinder treatment decision-making, and pose challenges for both clinical and translational research. This multi-institutional retrospective cohort study aimed to determine whether anatomic location or histologic subtype better predicts outcomes in AM. Distal extremity primary melanomas (n = 469) were matched 1:2 with cutaneous melanomas (CMs; n = 938). Cox proportional hazards models evaluated recurrence-free survival (RFS), melanoma-specific survival (MSS), and overall survival (OS). Of the distal tumors, 280 (59.7%) were ALMs, including 33 (11.8%) on non-subungual dorsal sites. Compared to CMs, acral surface melanomas had worse outcomes, while dorsal melanomas had similar outcomes, independent of histology. ALMs were associated with worse survival than superficial spreading melanomas (SSMs). In interaction models, ALMs on dorsal surfaces and all histologic subtypes on acral surfaces had worse MSS than non-distal SSMs. A revised definition of AM may better inform clinical management and future research.

BackgroundCovid-19 pandemic by the new coronavirus SARS-Cov-2 has produced devastating effects on the health care system, affecting also cancer patient care. Data about COVID-19 infection in cancer patients are scarce, and they point out a higher risk of complications due to the viral infection in this population. Moreover, cancer treatments could increase viral complications, specially those treatments based on the use of immunotherapy with checkpoints antibodies. There are no clinical data about the safety of immune check point antibodies in cancer patients when they become infected by SARS-CoV-2. As checkpoint inhibitors, mainly anti PD-1 and anti CTLA-4 antibodies, are an effective treatment for most melanoma patients, avoiding their use during the pandemic could lead to a decrease in the chances of curing melanoma. MethodsIn Spain we have started a national registry of melanoma patients infected by SARS-Cov-2 since April 1st, 2020. A retrospective analysis of patients included in the Spanish registery has been performed weekly since the activation of the study. Interim analysis shows unexpected findings about cancer treatment safety in SARS-Cov-2 infected melanoma patients, so a rapid communication to the scientific community is mandatory ResultsFifty patients have been included as of May 17th, 2020. Median age is 69 years (range 6 to 94 years), 27 (54%) patients are males and 36 (70%) patients have stage IV melanoma. Twenty-two (44%) patients were on active anticancer treatment with anti PD-1 antibodies, 16 (32%) patients were on treatment with BRAF plus MEK inhibitors and 12 (24%) patients were not on active cancer treatment. COVID-19 episode has been resolved in 43 cases, including 30 (70%) patients cured, four (9%) patients that have died due to melanoma progression, and nine (21%) patients that have died from COVID-19. Mortality rates from COVID-19 according to melanoma treatment type were 16%, 15% and 36% for patients on immunotherapy, targeted drugs, and for those that were not undergoing active cancer treatment, respectively. ConclusionThese preliminary findings show that the risk of death in those patients under going treatment with anti PD-1 antibodies does not exceed the global risk of death in this population. These results could be relevant in order to select melanoma therapy during the COVID-19 pandemic

PurposeDespite their central role in the current staging system, Breslow thickness and ulceration do not fully identify early-stage melanoma patients who will develop distant metastases. We assessed whether gene expression profiling (GEP) improves prediction of distant metastases beyond standard staging factors in early-stage melanoma. MethodsData were derived from the population-based Dutch Early-Stage Melanoma (D-ESMEL) study, including a matched discovery set of 442 stage I/II melanomas (221 case-control pairs) and a validation cohort of 308 melanomas nested within 5,815 patients. The discovery set was used to identify genes associated with distant metastases, independent of age, sex, Breslow thickness, and ulceration. The validation cohort was partitioned into model development and independent validation subsets. Candidate genes from the discovery set were used to develop and validate a GEP model, evaluated by weighted area under the curve (AUC) and concordance index (C-index). ResultsRNA sequencing succeeded for 356 melanomas in the discovery set, 200 in the model development subset, and 94 melanomas in the independent validation subset. Differential gene expression analyses and modeling identified 558 candidate genes. In the independent validation subset, the GEP model achieved a weighted AUC of 0.77 (95% CI, 0.66-0.86) and weighted C-index of 0.79 (95% CI, 0.69-0.88), comparable to the clinical model based on Breslow thickness and ulceration (weighted AUC 0.82 (95% CI, 0.73-0.90), weighted C-index 0.84 (95% CI, 0.76-0.91)). Integration of GEP with the clinical model did not improve accuracy. Gene set enrichment analyses showed enrichment of proliferative and stress-related pathways. ConclusionWhile GEP captured biologically relevant signals, its predictive accuracy for distant metastases was comparable to that of Breslow thickness and ulceration in a population-based early-stage melanoma cohort.

BackgroundCutaneous melanoma (CM) is an aggressive skin cancer with rising incidence, representing a growing public health concern. Despite the remarkable success of immune-checkpoint inhibitors (ICIs) in the management of advanced disease, mortality remains high due to therapy resistance. Identifying reliable prognostic and predictive biomarkers is therefore essential to improve patient stratification, optimize treatment selection, and minimize unnecessary toxicity. MethodsWe comprehensively profiled the circulating immune landscape of 54 treatment-naive CM patients by integrating flow cytometry immunophenotyping with clinicopathological data, and performed tumor gene expression analysis in a subset of 26 patients. ResultsElevated HLA-DR and CD69 expression on circulating CD4+ T cells, together with reduced circulating CD8+ T cell frequency, emerged as candidate prognostic biomarkers associated with improved survival. Prognostic models combining these immune variables with clinical covariates accurately stratified patients by overall survival (89.5% sensitivity, 72.7% specificity; AUC = 0.872, p < 0.0001) and progression/recurrence risk (75% sensitivity and 71.4% specificity; AUC = 0.763, p = 0.001). In a subset of 43 patients subsequently treated with ICIs, elevated baseline HLA-DR and CD69 expression on circulating CD4+ T cells was also associated with therapeutic benefit. A predictive model integrating these markers with clinical covariates achieved good discriminatory performance (65.2% sensitivity, 88.9% specificity; AUC = 0.775, p = 0.0027). Tumor gene expression profiling supported the role of IFN-{gamma}-related signatures, previously linked to ICI response, as complementary prognostic and predictive tools. ConclusionThese findings highlight systemic CD4+ T cell activation status as a promising, easily measurable biomarker in CM, laying the foundation for future strategies to refine patient stratification and guiding immunotherapy decisions.

Association of DNA hypomethylating agents (DHA) with immune-checkpoint inhibitors (ICI) is a promising strategy to improve efficacy of ICI-based therapy. Here we report the five-year clinical outcome and an integrated multi-omics analysis of pre- and on-treatment lesions from advanced melanoma patients enrolled in the phase Ib NIBIT-M4 study, a dose-escalation trial of the DHA agent guadecitabine combined with ipilimumab. With a minimum follow-up of 45 months the median OS was 25.6 months; the 5-year OS rate was 28.9% and the median DoR was 20.6 months. Specific genomic features and extent of T and B cellmediated immunity discriminated lesions of responding compared to non-responding patients. Enrichment for proliferation and EMT-related gene programs, and immune escape mechanisms characterized lesions from non-responding patients. Integration of a genetic immunoediting index (GIE) with an adaptive immunity signature (ICR) stratified patients/lesions into four distinct subsets and discriminated 5-year OS and PFS. These results suggest that coupling of immunoediting with activation of adaptive immunity is a relevant requisite for achieving long term clinical benefit by epigenetic immunomodulation in advanced melanoma patients.

Patients with advanced melanoma are treated with immune checkpoint inhibitors (ICIs), yet less than 50% of patients achieve a durable response while all patients are exposed to the risk of severe side effects. Tumor-infiltrating lymphocytes (TILs) in pathology images are associated with ICI outcomes, but manual assessment is subjective. In addition, the predictive value of other immune cell subsets, including plasma cells, neutrophils, histiocytes, and melanophages, remains unclear. We organized the Panoptic segmentation of nUclei and tissue in advanced MelanomA (PUMA) challenge to evaluate whether the spatial localization of TILs and other immune cell subsets on melanoma H&E slides collected before start of treatment was associated with treatment outcomes. Algorithm performance was evaluated on a hidden test set, after which top-ranked algorithms were applied to pre-treatment metastatic whole-slide images from a large, multicenter cohort of patients with advanced melanoma treated with first-line ICIs (n=1102). Automatically quantified tissue features and immune cell subsets were then associated with clinical outcomes. Top-performing algorithms improved detection of immune cell subsets, although accuracy for rare classes remained limited. Across challenge participants, TIL density showed the most consistent association with treatment response and survival. Associations for stromal TILs were weaker, while plasma cells, histiocytes, melanophages, neutrophils, necrosis and blood vessels did not show independent associations with outcomes. Overall, the results from the PUMA challenge improved the state of the art of immune cell detection in melanoma histopathology and show that intra-tumoral lymphocytes are the immune cell subset most consistently associated with treatment response and survival. HighlightsO_LIWe organized the first melanoma-specific tissue and nuclei segmentation competition C_LIO_LIWinning algorithms were applied to 1102 whole-slide images for biomarker analysis C_LIO_LIIntra-tumoral TILs were associated with response to immune checkpoint inhibitors C_LIO_LIOther immune cell subsets showed no independent association with treatment outcomes C_LIO_LITissue segmentation on WSIs was limited by low heterogeneity in training data. C_LI Graphical abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=140 SRC="FIGDIR/small/26347935v1_ufig1.gif" ALT="Figure 1"> View larger version (39K): org.highwire.dtl.DTLVardef@13838e4org.highwire.dtl.DTLVardef@1f34a6org.highwire.dtl.DTLVardef@b9a65borg.highwire.dtl.DTLVardef@58d300_HPS_FORMAT_FIGEXP M_FIG C_FIG

IntroductionUveal melanoma is a lethal intraocular tumour, in which the presence of certain genetic alterations correlates with the risk of metastatic dissemination and patient survival. RNA data is typically used to transcriptionally characterise tumours and their micro-environment. In this study, we tested the detectability of all key genetic alterations in uveal melanoma from RNA sequencing data. MethodsCohort-wide gene expression profiling was used to classify tumours at the transcriptional level. In individual samples, copy number alterations affecting chromosomes 3 and 8q were analysed by measuring expressed allelic imbalances of heterozygous common single nucleotide polymorphisms. Mutations in GNAQ, GNA11, CYSLTR2, PLCB4, BAP1, SF3B1 and EIF1AX were identified by screening of hotspot regions and by evaluating their transcriptional effects. All findings were cross-validated with DNA-derived data in a training cohort of 80 primary uveal melanomas studied by The Cancer Genome Atlas (TCGA) initiative, and in five prospectively analysed cases from our institution. ResultsUnsupervised gene expression profiling strongly correlated to the presence of chromosome 3 alterations, but was not reliable in identifying other (clinically-)relevant genetic alterations. However, the presence of both chromosome 3 and 8q copy number alterations could be successfully inferred from expressed allelic imbalances in most tumours. The majority of mutations were adequately recognised at the RNA level by their nucleotide changes (all genes), alternative splicing around the mutant position (BAP1) and transcriptome-wide aberrant splice junction usage (SF3B1). Notably, in the TCGA cohort we detected previously unreported mutations in BAP1 (n=3) and EIF1AX (n=5), that were missed by the original DNA sequencing. In our prospective cohort, all mutations and copy number alterations were successfully identified at the RNA level by combining the described approaches. ConclusionIn addition to providing gene expression levels and profiles, RNA from uveal melanomas presents insights into the expressed tumour genotype and its phenotypic consequences. Such complete analysis of transcriptional data may augment or even substitute current DNA-based approaches, and has potential applicability in both research and clinical practice.

IntroductionThe association of body composition with checkpoint inhibitor outcomes in melanoma is a matter of ongoing debate. In this study, we aim to add to previous evidence by investigating body mass index (BMI) alongside CT derived body composition metrics in the largest cohort to date. MethodPatients treated with first-line anti-PD1 {+/-} anti-CTLA4 for advanced melanoma were retrospectively identified from 11 melanoma reference centers in The Netherlands. Age, sex, Eastern Cooperative Oncology Group performance status, serum lactate dehydrogenase, presence of brain and liver metastases, number of affected organs and BMI at baseline were extracted from electronic patient files. From baseline CT scans, five body composition metrics were automatically extracted: skeletal muscle index, skeletal muscle density, skeletal muscle gauge, subcutaneous adipose tissue index and visceral adipose tissue index. All predictors were correlated in uni- and multivariable analysis to progression-free, overall and melanoma-specific survival (PFS, OS and MSS) using Cox proportional hazards models. ResultsA total of 1471 eligible patients were included. Median PFS and OS were 8.8 and 34.8 months, respectively. A significantly worse PFS was observed in underweight patients (multivariable HR=1.87, 95% CI 1.14-3.07). Furthermore, better OS was observed in patients with higher skeletal muscle density (multivariable HR=0.91, 95% CI 0.83-0.99) and gauge (multivariable HR=0.88, 95% CI 0.84-0.996), and a worse OS with higher visceral adipose tissue index (multivariable HR=1.13, 95% CI 1.04-1.22). No association with survival outcomes was found for overweightness or obesity and survival outcomes, or for subcutaneous adipose tissue. DiscussionOur findings suggest that underweight BMI is associated with worse PFS, whereas higher skeletal muscle density and lower visceral adipose tissue index were associated with better OS. These associations were independent of previously identified predictors, including sex, age, performance status and extent of disease. No significant association between higher BMI and survival outcomes was observed.

Cutaneous squamous cell carcinoma (cSCC) is one of the most common cancers in humans and kills as many people annually as melanoma. The mutational and transcriptional landscape of cSCC has identified driver mutations associated with disease progression as well as key pathway activation in the progression of pre-cancerous lesions. The understanding of the transcriptional changes with respect to high-risk clinical/histopathological features and outcome is poor. Here, we examine stage-matched, outcome-differentiated cSCC and associated clinicopathologic risk factors using whole exome and transcriptome sequencing on matched samples. Exome analysis identified key driver mutations including TP53, CDKN2A, NOTCH1, SHC4, MIIP, CNOT1, C17orf66, LPHN22, and TTC16 and pathway enrichment of driver mutations in replicative senescence, cellular response to UV, cell-cell adhesion, and cell cycle. Transcriptomic analysis identified pathway enrichment of immune signaling/inflammation, cell-cycle pathways, extracellular matrix function, and chromatin function. Our integrative analysis identified 183 critical genes in carcinogenesis and were used to develop a gene expression panel (GEP) model for cSCC. Three outcome-related gene clusters included those involved in keratinization, cell division, and metabolism. We found 16 genes were predictive of metastasis (Risk score [&ge;] 9 Met & Risk score < 9 NoMet). The Risk score has an AUC of 97.1% (95% CI: 93.5% - 100%), sensitivity 95.5%, specificity 85.7%, and overall accuracy of 90%. Eleven genes were chosen to generate the risk score for Overall Survival (OS). The Harrells C-statistic to predict OS is 80.8%. With each risk score increase, the risk of death increases by 2.47 (HR: 2.47, 95% CI: 1.64-3.74; p<0.001) after adjusting for age, immunosuppressant use, and metastasis status.

Melanoma remains a formidable clinical challenge, claiming the lives of 60,000 patients annually. Current therapeutic modalities encompass surgical intervention and immune checkpoint inhibitors (ICB). Nevertheless, the efficacy of ICB varies, necessitating the need to anticipate response and resistance outcomes, while also considering alternative approaches, such as surgical interventions coupled with autologous skin grafts. In pursuit of these objectives, we conducted a comprehensive analysis involving seven melanoma patient cohorts subjected to four distinct ICB treatments. Remarkably, our findings revealed varying response rates: 29% for Nivolumab, 43% for Pembrolizumab, 20% for Ipilimumab, and an encouraging 62.5% for the combination of Pembrolizumab and Ipilimumab. This underscores the superior clinical outcomes associated with anti-PD1+anti-CTLA4 therapy. Intriguingly, responders to Pembrolizumab and Ipilimumab exhibited distinct immunological characteristics, characterized by an augmentation in Th1 and M1 macrophages, alongside a reduction in CD4+ T cell infiltration. This phenomenon coincided with the upregulation of antigen presentation genes (HLA, CD80), heightened pro-inflammatory cytokine production (CCL5, CXCL9, CXCL10), and enhanced T cell responses. Furthermore, based on these response profiles, we have developed predictive software to forecast individual patient responses to available checkpoint inhibitor combinations. This innovative tool also facilitates precise calculations for the extent of melanoma resection required during surgery, graft sizing, and the determination of the necessary autologous skin cell resources. In conclusion, our approach advocates for tailored therapies, leveraging patient-specific attributes and computational predictions to enhance clinical outcomes following immunotherapy and surgical interventions. This strategy holds promise for advancing melanoma treatment paradigms.

PurposeSolid tumors harboring tumor mutational burden (TMB) [&ge;] 10 mutations per megabase (mut/Mb) received agnostic approval for pembrolizumab. However, TMB cut-off alone is not a predictor of overall survival (OS). This work aims to analyze the somatic mutational profiles influence in the outcomes of patients with TMB-high tumors treated with immune checkpoint inhibitors (ICIs). MethodsThis post-hoc analysis evaluated clinical and molecular features of 1,661 patients with solid tumors treated with ICIs. We performed OS analysis for TMB thresholds of [&ge;] 10, [&ge;] 20, and < 10 mut/Mb. For a TMB [&ge;] 10mut/Mb cutoff, we assessed OS according to mutational profile. For genes exhibiting a correlation with OS (P < 0.05) at the univariate assessment, we conducted a Cox multivariate analysis adjusted by median TMB, sex, age, microsatellite instability (MSI), and histology. Results1,661 patients were investigated, and 488 harbored a TMB [&ge;] 10 mut/Mb (29.4%). The median OS was 42 months for TMB [&ge;] 10 or 20 mut/Mb, and 15 months for TMB < 10 mut/Mb (P < 0.005). In patients harboring TMB [&ge;] 10mut/Mb, mutations in E2F3 or STK11 were correlated with worse OS, and mutations in NTRK3, PTPRD, RNF43, TENT5C, TET1 or ZFHX3 with better OS. These associations were confirmed by univariate and multivariate analyses (P < 0.05). Melanoma histology and TMB above the median endowed patients with better OS (P < 0.05). MSI status, age, and gender did not have a consistent statistically significant effect on OS ConclusionCombining TMB information and mutation profiles in key cancer genes can be used to better qualify patients for ICI treatment and predict their OS. CONTEXT SUMMARYO_ST_ABSKey objectiveC_ST_ABSTumor mutational burden (TMB) of [&ge;] 10 mutations per megabase (mut/Mb) grant agnostic indication of pembrolizumab for advanced solid tumors treatment, however a substantial number of patients do not respond to therapy. This work aims to analyze the somatic mutational profiles influence in the outcomes of patients with TMB [&ge;] 10mut/Mb tumors treated with ICIs. Knowledge generatedMutation profile can modify survival outcomes to ICIs in patients with TMB [&ge;] 10 mut/Mb. Mutations in E2F3 or STK11 correlate with worse OS, while mutations in NTRK3, PTPRD, RNF43, TENT5C, TET1 or ZFHX3 correlate with better OS in TMB-high patients receiving ICIs. RelevanceWe found that the combination of a high TMB and the somatic mutational profile in key cancer genes can be decisive in better qualify patients for ICI treatment.

BackgroundLarge language models (LLMs) are gaining recognition across various medical fields; however, their specific role in dermatology, particularly in melanoma care, is not well- defined. This systematic review evaluates the current applications, advantages, and challenges associated with the use of LLMs in melanoma care. MethodsWe conducted a systematic search of PubMed and Scopus databases for studies published up to July 23, 2024, focusing on the application of LLMs in melanoma. Identified studies were categorized into three subgroups: patient education, diagnosis and clinical management. The review process adhered to PRISMA guidelines, and the risk of bias was assessed using the modified QUADAS-2 tool. ResultsNine studies met the inclusion criteria. Five studies compared various LLM models, while four focused on ChatGPT. Three studies specifically examined multi-modal LLMs. In the realm of patient education, ChatGPT demonstrated high accuracy, though it often surpassed the recommended readability levels for patient comprehension. In diagnosis applications, multi- modal LLMs like GPT-4V showed capabilities in distinguishing melanoma from benign lesions. However, the diagnostic accuracy varied considerably, influenced by factors such as the quality and diversity of training data, image resolution, and the models ability to integrate clinical context. Regarding management advice, one study found that ChatGPT provided more reliable management advice compared to other LLMs, yet all models lacked depth and specificity for individualized decision-making. ConclusionsLLMs, particularly multimodal models, show potential in improving melanoma care through patient education, diagnosis, and management advice. However, current LLM applications require further refinement and validation to confirm their clinical utility. Future studies should explore fine-tuning these models on large dermatological databases and incorporate expert knowledge.

IntroductionPredicting checkpoint inhibitors treatment outcomes in melanoma is a relevant task, due to the unpredictable and potentially fatal toxicity and high costs for society. However, accurate biomarkers for treatment outcomes are lacking. Radiomics are a technique to quantitatively capture tumor characteristics on readily available computed tomography (CT) imaging. The purpose of this study was to investigate the added value of radiomics for predicting durable clinical benefit from checkpoint inhibitors in melanoma in a large, multicenter cohort. MethodsPatients who received first-line anti-PD1 {+/-} anti-CTLA4 treatment for advanced cutaneous melanoma were retrospectively identified from nine participating hospitals. For every patient, up to five representative lesions were segmented on baseline CT and radiomics features were extracted. A machine learning pipeline was trained on the radiomics features to predict durable clinical benefit, defined as stable disease for more than six months or response per RECIST 1.1 criteria. This approach was evaluated using a leave-one-center-out cross validation and compared to a model based on previously discovered clinical predictors. Lastly, a combination model was built on the radiomics and clinical model. ResultsA total of 620 patients were included, of which 59.2% experienced durable clinical benefit. The radiomics model achieved an area under the receiver operator characteristic curve (AUROC) of 0.607 [95%CI 0.562-0.652], lower than that of the clinical model (AUROC=0.646 [95%CI 0.600-0.692]). The combination model yielded no improvement over the clinical model in terms of discrimination (AUROC=0.636 [95%CI 0.592-0.680]) or calibration. The output of the radiomics model was significantly correlated with three out of five input variables of the clinical model (p < 0.001). DiscussionThe radiomics model achieved a moderate predictive value of durable clinical benefit, which was statistically significant. However, a radiomics approach was unable to add value to a simpler clinical model, most likely due to the overlap in predictive information learned by both models. Future research should focus on the application of deep learning, spectral CT derived radiomics and a multimodal approach for accurately predicting benefit to checkpoint inhibitor treatment in advanced melanoma.

BackgroundCutaneous squamous cell carcinoma (cSCC) is the second most common non-melanoma skin cancer and is associated with considerable morbidity. Population-based data analysis in Germany has largely focused on incidence and trends. ObjectivesTo assess incidence, anatomical site- and T-stage distribution, histological subtypes of cSCC in Germany, with a focus on sex- and age-group specific patterns and regional differences. MethodsA total of 213,935 first primary invasive cSCC cases diagnosed between 1986 and 2019 were analysed from four federal states of Germany with complete case ascertainment. Crude and age-standardized incidence rates (CIR, ASIR) were calculated, and subgroup analyses were performed by sex, anatomical site, histological subtype, and T-stage and region. ResultsCIR increased by over 500 % from 1986 to 2015, with a steeper rise in women. Incidence plateaued after 2015 in most states, except for a delayed increase in Saarland. The face (ICD-10 C44.3) was the most frequent tumor site showing equal incidence in males and females. T1 tumors predominated (88.6 %), although staging data were incomplete in 33.5 % of cases. Regional and sex-based differences were observed in both T-stage and histological subtype distribution. Spindle cell and non-keratinizing variants were associated with more advanced stages. Cancer registry data did not count more than one cSCC and carcinoma in situ such as Bowens disease or actinic keratosis, leading to systematic underestimation of disease burden. ConclusionscSCC incidence has risen substantially in Germany, with significant variation by sex, region, and tumor type. Improved registry protocols incorporating multiple primaries, clinical staging, and early in situ lesions are essential for accurate surveillance and healthcare planning. Plain Language SummaryHow common is squamous cell skin cancer in Germany and how does it behave? We looked at cutaneous squamous cell carcinoma (cSCC), a common skin cancer that starts in the flat cells on the skins surface. It is the second most common skin cancer and the second most common cancer, affecting tens of thousands of people in Germany each year. We found that the registration of new primary cSCC tumors increased more than fivefold between 1986 and 2015. However, incidence rates plateaued from 2015 to 2019. Tumors most often appeared on the face, but the distribution by site differed between men and women. Men developed cSCC at younger ages and more frequently than women. Approximately 90% of tumors were diagnosed at an early stage, but staging information was missing for about 34% of cases. cSCC develops on chronically sun-damaged skin, and patients often have more than one tumor. There are also early skin changes that require treatment. Because cancer registries count only one tumor per person and ignore these early lesions, the true burden of treating patients with chronic sun damage is underestimated. We concluded that cSCC has become more common in Germany, with clear differences by sex, age, and region. Improving cancer registries to record all tumors will provide a more accurate picture of stage and subtype distribution. Recognizing high-risk groups, will help guide prevention, screening, and earlier treatment strategies. What is already known about this topic?Cutaneous squamous cell carcinoma (cSCC) is the second most frequent malignancy overall as well as the second most frequent skin tumor. Epidemiological research has long focused on Basal Cell Carcinoma (BCC) and cSCC conclusively. Recent research has addressed major differences in their epidemiology, highlighting differnces in incidence rates across genders and age groups. The largest data set on squamous cell carcinoma analysed so far was 145.000 cases. What does this study add?This study provides conclusive results on incidence, localization, tumor stages and histologic types comparing men, women and age groups based on large scale data with over 200,000 primary tumors over a period of more than 30 years. What is the translational message?Identification of gender- and age-specific risk patterns in cSCC enables the formulation of targeted prevention strategies, screening recommendations, and earlier diagnosis and optimized management of high-risk populations. The burden of morbidity and tumors associated with chronic actinic damage remains underestimated in current literature and cancer statistics. Demographic changes are expected to increase the disease burden substantially, although the exact magnitude remains uncertain.

The phase 1b TICIMEL clinical trial evaluated the safety, tolerability, and anti-tumor activity of combining the immune checkpoint inhibitors (ICI), ipilimumab and nivolumab, with tumor necrosis factor (TNF) blockers, certolizumab or infliximab, to treat advanced melanoma patients. A higher proportion of responses was observed in patients receiving ICI and certolizumab, while patients treated with ICI and infliximab demonstrated superior tolerability. Moreover, CITE-Seq analyses of circulating CD8 T cells showed that ICI plus certolizumab promoted an IFN signature, whereas ICI plus infliximab reduced the induction of genes associated with T cell activation. In preclinical models, ICI and TNF blockade with certolizumab increased IFN-{gamma}+ CD8 T cells and reduced regulatory T cells in tumors. The IgG1 Fc fragment of infliximab was identified as counteracting the benefits of TNF blockade. These findings underscore the importance of selecting the optimal TNF blocker to combine with ICI to enhance therapy efficacy in melanoma patients. ClinicalTrials.gov identifiers: NCT03293784; NCT05867004.

BackgroundAdoptive cell transfer (ACT) of Tumour-infiltrating lymphocytes (TIL) is an investigational treatment for solid tumours, with preliminary results showing objective clinical responses in some metastatic melanoma patients. The ability to sequence and track the T cell repertoire throughout ACT of TILs provides a method to identify T cell repertoire features associated with patients benefit from ACT. Identification of response biomarkers for patients receiving ACT of TILs has been limited. Conflicting evidence is observed in biomarkers such as the number of TILs in the infusion product, with some studies suggesting a relationship with response and others not. Meanwhile, certain potential biomarkers, such as the diversity of the post- infusion peripheral repertoire, have not yet been studied. MethodsIn this study, we sought to determine 1) the efficacy of using CapTCR-seq to track TILs in serial blood draws over the course of ACT immunotherapy 2) whether peripheral T cell repertoire statistics are associated with ACT response. In this study, 9 patients with cutaneous (n = 7) or mucosal (n = 2) melanoma received TIL ACT after chemotherapeutic depletion. Hybrid- capture CapTCR-seq was conducted on pre-/post-transfer peripheral blood mononuclear cells (PBMC) and cell-free DNA. ResultsComparison between PBMC DNA, PBMC RNA, and circulating free DNA (cfDNA) repertoires demonstrated an increased presence of shared T cell clonotypes post-infusion when compared with baseline samples. Higher abundance of TIL clonotypes in the PBMC baseline and post-infusion DNA T cell repertoires and the presence of shared DNA T cell clonotypes between timepoints was seen in responders when compared with non-responders according to RECIST criteria. ConclusionThese results demonstrate effective tracking methodologies and suggest a predictive role for baseline repertoire statistics in response to the ACT of TILs. Trial registrationUniversity Health Network Research Ethics Board #11-0683

Quantitative assessment of multiplex immunofluorescence (mIF) data represents a powerful tool for immunotherapy biomarker discovery in melanoma and other solid tumors. In addition to providing detailed phenotypic information of immune cells of the tumor microenvironment, these datasets contain spatial information that can reveal biologically relevant interactions among cell types. To assess quantitative mIF analysis as a platform for biomarker discovery, we used a 12-plex mIF panel to characterize tumor samples collected from 50 patients with melanoma prior to treatment with immune checkpoint inhibitors (ICI). Consistent with prior studies, we identified a strong association between stromal B cell percentage and response to ICI therapy. We then compared pathologist assessment of lymphoid aggregates with a density based clustering algorithm, DBSCAN, to both automatically detect B cell aggregates and quantify their size, morphology, and distance to tumor. Spatial neighborhood analysis identified TCF1+ and LAG3-T cell subpopulations enriched near stromal B cells. These analyses provide a roadmap for the further development and validation of spatial immunotherapy biomarkers in melanoma and other diseases.

BackgroundMelanoma is a lethal form of skin cancer with a high propensity for metastasizing, making early detection crucial. This study aims to develop a machine learning model using electronic health record data to identify patients at high risk of developing melanoma to prioritize them for dermatology screening. MethodsThis retrospective study included patients diagnosed with melanoma (cases), as well as matched patients without melanoma (controls), from Massachusetts General Hospital (MGH), Brigham and Womens Hospital (BWH), Dana-Farber Cancer Institute (DFCI), and other hospital centers within the Research Patient Data Registry at Mass General Brigham healthcare system between 1992 and 2022. Patient demographics, family history, diagnoses, medications, procedures, laboratory tests, reasons for visits, and allergy data six months prior to the date of first melanoma diagnosis or date of censoring were extracted. A machine learning framework for health outcomes (MLHO) was utilized to build the model. Performance was evaluated using five-fold cross-validation of the MGH cohort (internal validation) and by using the MGH cohort for model training and the non-MGH cohort for independent testing (external validation). The Area Under the Receiver Operating Characteristic Curve (AUC-ROC) and the Area Under the Precision-Recall Curve (AUC-PR), along with 95% Confidence Intervals (CIs), were computed. ResultsThis study identified 10,778 patients with melanoma and 10,778 matched patients without melanoma, including 8,944 from MGH and 1,834 from non-MGH hospitals in each cohort, both with an average follow-up duration of 9 years. In the internal and external validations, the model achieved AUC-ROC values of 0.826 (95% CI: 0.819-0.832) and 0.823 (95% CI: 0.809-0.837) and AUC-PR scores of 0.841 (95% CI: 0.834-0.848) and 0.822 (95% CI: 0.806-0.839), respectively. Important risk features included a family history of melanoma, a family history of skin cancer, and a prior diagnosis of benign neoplasm of skin. Conversely, medical examination without abnormal findings was identified as a protective feature. ConclusionsMachine learning techniques and electronic health records can be effectively used to predict melanoma risk, potentially aiding in identifying high-risk patients and enabling individualized screening strategies for melanoma.