Outcomes of patients with resected stage III/IV acral or mucosal melanoma, treated with adjuvant anti-PD-1 based therapy.

IMPORTANCE: Acral (AM) and mucosal melanomas (MM) are rare subtypes with a poor prognosis. In those with advanced disease, anti-PD-1 (PD1) therapy has reduced activity compared to that seen in non-acral cutaneous melanoma. OBJECTIVE: To determine the efficacy of adjuvant PD1 in resected AM or MM. DESIGN: An international, retrospective cohort study SETTING: Data up to November 2021 collected from 20 centres across 10 countries. PARTICIPANTS: One hundred and ninety four patients with resected stage III or IV1 AM or MM who received adjuvant PD1 were included and compared to matched patients from the Melanoma Institute Australia (MIA) database using a propensity score matching analysis. MAIN OUTCOMES AND MEASURES: Recurrence-free survival (RFS), distant metastasis-free survival (DMFS) and overall survival (OS) were investigated. RESULTS: Forty five of 139 (32%) AM and 9 of 55 (16%) MM patients completed adjuvant therapy. The main reason for early treatment cessation in both groups was disease recurrence: 51 (37%) and 30 (55%) in the AM and MM groups, respectively. In the AM group adjuvant PD1 was associated with a longer RFS [HR-0.69 (0.52-0.92, p = 0.0127)], DMFS [HR0.58 (0.38-0.89, p = 0.0134)] and OS [HR of 0.59 (0.38-0.92, p-value 0.0196)] when compared to the historical cohort. In the MM group there was no statistical difference in RFS [HR1.36 (0.69-2.68,p-value 0.3799], DMFS or OS. CONCLUSION AND RELEVANCE: After adjuvant PD1, both AM and MM have a high risk of recurrence. Our data suggests a benefit to using adjuvant PD1 therapy in resected AM but not in resected MM. Additional studies to investigate the efficacy of adjuvant PD1 for MM are needed.

Efficacy and safety of 'Second Adjuvant' therapy with BRAF/MEK inhibitors after local therapy for recurrent melanoma following adjuvant PD-1 based immunotherapy.

BACKGROUND: Anti-PD-1 antibodies and BRAK/MEK inhibitors (BRAF/MEKi) reduce the risk of recurrence for patients with resected stage III melanoma. BRAFV600-mutated (BRAFmut) melanoma patients who recur with isolated disease following adjuvant therapy may be suitable for 'second adjuvant' treatment after local therapy. We sought to examine the efficacy and safety of 'second adjuvant' BRAF/MEKi. PATIENTS AND METHODS: Patients with BRAFmut melanoma treated with adjuvant PD-1 based immunotherapy who recurred, underwent definitive local therapy and were then treated with adjuvant BRAF/MEKi were identified retrospectively from 13 centres (second adjuvant group). Demographics, disease and treatment characteristics and outcome data were examined. Outcomes were compared to BRAFmut patients who did not receive 'second adjuvant' therapy (no second adjuvant group). RESULTS: 73 patients were included; 61 who received 'second adjuvant' therapy and 12 who did not. Most initially recurred on PD-1 therapy (66%). There were no differences in characteristics between groups. 92% of second adjuvant group received dabrafenib and trametinib and median duration of therapy was 11.8 months (0.4, 34.5). 72% required dose adjustments, 23% had grade 3 + toxicity and 38% permanently discontinued drug due to toxicity. After median 26.1 months (1.9, 56.3) follow-up, recurrence-free survival (RFS) was improved in second adjuvant group versus no second adjuvant group (median 30.8 vs 4 months, HR 0.35; p = 0.014), largely driven by a delay in early recurrence, with no difference in overall survival (p = 0.59). CONCLUSION: This is the first study examining outcomes of 'second adjuvant' targeted therapy for melanoma, after failure of adjuvant PD-1 based immunotherapy. Data suggest a short-term improvement in RFS, but at the cost of toxicity. Alternative strategies and more data on sequencing adjuvant therapies are required to improve outcomes.

The role of diabetes in metastatic melanoma patients treated with nivolumab plus relatlimab.

BACKGROUND: The combination of nivolumab + relatlimab is superior to nivolumab alone in the treatment of naive patients and has activity in PD-1 refractory melanoma. We had previously observed a reduced expression of LAG3 in melanoma tissue from patients with type 2 diabetes. METHOD: To evaluate the impact of diabetes on oncological outcomes of patients with advanced melanoma treated with nivolumab plus the LAG3 inhibitor relatlimab we performed a retrospective multicenter study. RESULTS: Overall, 129 patients were included: 88 without diabetes before the treatment, 37 who were diagnosed with type 2 diabetes before the start of treatment, and 4 without diabetes before treatment who developed immune checkpoint inhibitor-induced diabetes (ICI-DM). PFS was 21.71 months (95% CI: 15.61-27.81) in patients without diabetes, 10.23 months (95% CI: 5.81-14.66) in patients with type 2 diabetes, and 50.85 months (95% CI: 23.04-78.65) in patients who developed ICI-DM. OS was 37.94 months (95% CI: 31.02-44.85) in patients without diabetes, 22.12 months (95% CI: 14.41-29.85) in those with type 2 diabetes and 57.64 months (95% CI: 42.29-72.99) in those who developed ICI-DM. Multivariate analysis showed that the presence of diabetes and LDH was correlated with OS and PFS. The mean OS was 64.63 months in subjects with low levels of glucose (< 137 mg/dl) and 36.27 months in those with high levels (hazard ratio 0.16, 95% CI: 0.04-0.58; p = 0.005). The patients whose glucose blood level increased after 3 months of treatment with nivolumab + relatinib compared to baseline (ratio of blood level at baseline/after 3 months > 1.5) had a worse prognosis than those whose glucose level had not increased. This result was observed also in subgroups treated either in first line or further lines. Patients who developed ICI-DM during the study period had better outcomes than the overall population and patients without diabetes. CONCLUSIONS: LAG3 inhibition for treating metastatic or unresectable melanoma has a reduced efficacy in patients with type 2 diabetes, possibly due to a low expression of LAG3 in tumor tissue. Higher level evidence should be obtained.

Modeling tumor size dynamics based on real-world electronic health records and image data in advanced melanoma patients receiving immunotherapy.

The development of immune checkpoint inhibitors (ICIs) has revolutionized cancer therapy but only a fraction of patients benefits from this therapy. Model-informed drug development can be used to assess prognostic and predictive clinical factors or biomarkers associated with treatment response. Most pharmacometric models have thus far been developed using data from randomized clinical trials, and further studies are needed to translate their findings into the real-world setting. We developed a tumor growth inhibition model based on real-world clinical and imaging data in a population of 91 advanced melanoma patients receiving ICIs (i.e., ipilimumab, nivolumab, and pembrolizumab). Drug effect was modeled as an ON/OFF treatment effect, with a tumor killing rate constant identical for the three drugs. Significant and clinically relevant covariate effects of albumin, neutrophil to lymphocyte ratio, and Eastern Cooperative Oncology Group (ECOG) performance status were identified on the baseline tumor volume parameter, as well as NRAS mutation on tumor growth rate constant using standard pharmacometric approaches. In a population subgroup (n = 38), we had the opportunity to conduct an exploratory analysis of image-based covariates (i.e., radiomics features), by combining machine learning and conventional pharmacometric covariate selection approaches. Overall, we demonstrated an innovative pipeline for longitudinal analyses of clinical and imaging RWD with a high-dimensional covariate selection method that enabled the identification of factors associated with tumor dynamics. This study also provides a proof of concept for using radiomics features as model covariates.

Multilingual RECIST classification of radiology reports using supervised learning.

Objectives: The objective of this study is the exploration of Artificial Intelligence and Natural Language Processing techniques to support the automatic assignment of the four Response Evaluation Criteria in Solid Tumors (RECIST) scales based on radiology reports. We also aim at evaluating how languages and institutional specificities of Swiss teaching hospitals are likely to affect the quality of the classification in French and German languages. Methods: In our approach, 7 machine learning methods were evaluated to establish a strong baseline. Then, robust models were built, fine-tuned according to the language (French and German), and compared with the expert annotation. Results: The best strategies yield average F1-scores of 90% and 86% respectively for the 2-classes (Progressive/Non-progressive) and the 4-classes (Progressive Disease, Stable Disease, Partial Response, Complete Response) RECIST classification tasks. Conclusions: These results are competitive with the manual labeling as measured by Matthew's correlation coefficient and Cohen's Kappa (79% and 76%). On this basis, we confirm the capacity of specific models to generalize on new unseen data and we assess the impact of using Pre-trained Language Models (PLMs) on the accuracy of the classifiers.

Semiautomated Pipeline to Quantify Tumor Evolution From Real-World Positron Emission Tomography/Computed Tomography Imaging.

PURPOSE: A semiautomated pipeline for the collection and curation of free-text and imaging real-world data (RWD) was developed to quantify cancer treatment outcomes in large-scale retrospective real-world studies. The objectives of this article are to illustrate the challenges of RWD extraction, to demonstrate approaches for quality assurance, and to showcase the potential of RWD for precision oncology. METHODS: We collected data from patients with advanced melanoma receiving immune checkpoint inhibitors at the Lausanne University Hospital. Cohort selection relied on semantically annotated electronic health records and was validated using process mining. The selected imaging examinations were segmented using an automatic commercial software prototype. A postprocessing algorithm enabled longitudinal lesion identification across imaging time points and consensus malignancy status prediction. Resulting data quality was evaluated against expert-annotated ground-truth and clinical outcomes obtained from radiology reports. RESULTS: The cohort included 108 patients with melanoma and 465 imaging examinations (median, 3; range, 1-15 per patient). Process mining was used to assess clinical data quality and revealed the diversity of care pathways encountered in a real-world setting. Longitudinal postprocessing greatly improved the consistency of image-derived data compared with single time point segmentation results (classification precision increased from 53% to 86%). Image-derived progression-free survival resulting from postprocessing was comparable with the manually curated clinical reference (median survival of 286 v 336 days, P = .89). CONCLUSION: We presented a general pipeline for the collection and curation of text- and image-based RWD, together with specific strategies to improve reliability. We showed that the resulting disease progression measures match reference clinical assessments at the cohort level, indicating that this strategy has the potential to unlock large amounts of actionable retrospective real-world evidence from clinical records.

Interactive process mining of cancer treatment sequences with melanoma real-world data.

The growing availability of clinical real-world data (RWD) represents a formidable opportunity to complement evidence from randomized clinical trials and observe how oncological treatments perform in real-life conditions. In particular, RWD can provide insights on questions for which no clinical trials exist, such as comparing outcomes from different sequences of treatments. To this end, process mining is a particularly suitable methodology for analyzing different treatment paths and their associated outcomes. Here, we describe an implementation of process mining algorithms directly within our hospital information system with an interactive application that allows oncologists to compare sequences of treatments in terms of overall survival, progression-free survival and best overall response. As an application example, we first performed a RWD descriptive analysis of 303 patients with advanced melanoma and reproduced findings observed in two notorious clinical trials: CheckMate-067 and DREAMseq. Then, we explored the outcomes of an immune-checkpoint inhibitor rechallenge after a first progression on immunotherapy versus switching to a BRAF targeted treatment. By using interactive process-oriented RWD analysis, we observed that patients still derive long-term survival benefits from immune-checkpoint inhibitors rechallenge, which could have direct implications on treatment guidelines for patients able to carry on immune-checkpoint therapy, if confirmed by external RWD and randomized clinical trials. Overall, our results highlight how an interactive implementation of process mining can lead to clinically relevant insights from RWD with a framework that can be ported to other centers or networks of centers.

A differential process mining analysis of COVID-19 management for cancer patients.

During the acute phase of the COVID-19 pandemic, hospitals faced a challenge to manage patients, especially those with other comorbidities and medical needs, such as cancer patients. Here, we use Process Mining to analyze real-world therapeutic pathways in a cohort of 1182 cancer patients of the Lausanne University Hospital following COVID-19 infection. The algorithm builds trees representing sequences of coarse-grained events such as Home, Hospitalization, Intensive Care and Death. The same trees can also show probability of death or time-to-event statistics in each node. We introduce a new tool, called Differential Process Mining, which enables comparison of two patient strata in each node of the tree, in terms of hits and death rate, together with a statistical significance test. We thus compare management of COVID-19 patients with an active cancer in the first vs. second COVID-19 waves to quantify hospital adaptation to the pandemic. We also compare patients having undergone systemic therapy within 1 year to the rest of the cohort to understand the impact of an active cancer and/or its treatment on COVID-19 outcome. This study demonstrates the value of Process Mining to analyze complex event-based real-world data and generate hypotheses on hospital resource management or on clinical patient care.

Systematic comparison with autoimmune liver disease identifies specific histological features of immune checkpoint inhibitor-related adverse events.

BACKGROUND: Immune checkpoint inhibitors (ICIs) have become a mainstay of cancer treatment. Their immune-boosting quality has one major drawback, their proclivity to induce a broad array of immune-related adverse events (irAEs) affecting, among others, the liver and sharing some similarities with classic autoimmune liver diseases (AILD).We aimed to compare clinical, laboratory and histological features of patients with liver-related irAEs and AILD. METHODS: We systematically compared liver irAEs with AILD, namely autoimmune hepatitis (AIH) and primary biliary cholangitis, regarding their clinical, laboratory, and histological features. RESULTS: Twenty-seven patients with liver irAEs (ICI group) and 14 patients with AILD were identified. We observed three distinct ICI-induced histological liver injury patterns: hepatitic (52%), cholangitic (19%), and mixed (29%). When comparing the ICI and AILD groups, centrilobular injury as well as granuloma formation were more prevalent in the former (p=0.067 and 0.002, respectively). CD4+/CD8+ T cell ratios were heterogeneous between the two groups, without statistically significant difference but with a trend toward increased CD8+ T cells among hepatitic irAEs as compared with AIH. Pattern of liver function test alteration was predictive for the type of irAEs but did not correlate with histological severity. CONCLUSIONS: Liver irAEs have broad clinical, laboratory and histological presentations. Histological features of irAEs and AILD are distinct, likely underpinning their different immunological mechanisms.

Efficacy of anti-PD-1 and ipilimumab alone or in combination in acral melanoma.

BACKGROUND: Acral melanoma is a rare melanoma subtype with poor prognosis. Importantly, these patients were not identified as a specific subgroup in the landmark melanoma trials involving ipilimumab and the anti-programmed cell death protein-1 (PD-1) agents nivolumab and pembrolizumab. There is therefore an absence of prospective clinical trial evidence regarding the efficacy of checkpoint inhibitors (CPIs) in this population. Acral melanoma has lower tumor mutation burden (TMB) than other cutaneous sites, and primary site is associated with differences in TMB. However the impact of this on the effectiveness of immune CPIs is unknown. We examined the efficacy of CPIs in acral melanoma, including by primary site. METHODS: Patients with unresectable stage III/IV acral melanoma treated with CPI (anti-PD-1 and/or ipilimumab) were studied. Multivariable logistic and Cox regression analyses were conducted. Primary outcome was objective response rate (ORR); secondary outcomes were progression-free survival (PFS) and overall survival (OS). RESULTS: In total, 325 patients were included: 234 (72%) plantar, 69 (21%) subungual and 22 (7%) palmar primary sites. First CPI included: 184 (57%) anti-PD-1, 59 (18%) anti-PD-1/ipilimumab combination and 82 (25%) ipilimumab. ORR was significantly higher with initial anti-PD-1/ipilimumab compared with anti-PD-1 (43% vs 26%, HR 2.14, p=0.0004) and significantly lower with ipilimumab (15% vs 26%, HR 0.49, p=0.0016). Landmark PFS at 1 year was highest for anti-PD-1/ipilimumab at 34% (95% CI 24% to 49%), compared with 26% (95% CI 20% to 33%) with anti-PD-1 and 10% (95% CI 5% to 19%) with ipilimumab. Despite a trend for increased PFS, anti-PD-1/ipilimumab combination did not significantly improve PFS (HR 0.85, p=0.35) or OS over anti-PD-1 (HR 1.30, p=0.16), potentially due to subsequent therapies and high rates of acquired resistance. No outcome differences were found between primary sites. CONCLUSION: While the ORR to anti-PD-1/ipilimumab was significantly higher than anti-PD-1 and PFS numerically higher, in this retrospective cohort this benefit did not translate to improved OS. Future trials should specifically include patients with acral melanoma, to help determine the optimal management of this important melanoma subtype.

Estimated Costs of the Ipilimumab-Nivolumab Therapy and Related Adverse Events in Metastatic Melanoma.

Combined ipilimumab and nivolumab significantly improve outcomes in metastatic melanoma patients but bear an important financial impact on the healthcare system. Here, we analyze the treatment costs, focusing on irAE. We conducted a retrospective analysis of 62 melanoma patients treated with ipilimumab-nivolumab at the Lausanne University Hospital between 1 June 2016 and 31 August 2019. The frequency of irAEs and outcomes were evaluated. All melanoma-specific costs were analyzed from the first ipilimumab-nivolumab dose until the therapy given subsequently or death. A total of 54/62 (87%) patients presented at least one irAE, and 31/62 (50%) presented a grade 3-4 irAE. The majority of patients who had a complete response 12/14 (86%) and 21/28 (75%) of overall responders presented a grade 3-4 toxicity, and there were no responses in patients without toxicity. Toxicity costs represented only 3% of the total expenses per patient. The most significant contributions were medication costs (44%) and disease costs (39%), mainly disease-related hospitalization costs, not toxicity-related. Patients with a complete response had the lowest global median cost per week of follow up (EUR 2425) and patients who had progressive disease (PD), the highest one (EUR 8325). Except for one patient who had a Grade 5 toxicity (EUR 6043/week), we observe that less severe toxicity grades (EUR 9383/week for Grade 1), or even the absence of toxicity (EUR 9922/week), are associated with higher median costs per week (vs. EUR 3266/week for Grade 4 and EUR 2850/week for Grade 3). The cost of toxicities was unexpectedly low compared to the total costs, especially medication costs. Patients with higher toxicity grades had better outcomes and lower total costs due to treatment discontinuation.

Myeloid antigen-presenting cell niches sustain antitumor T cells and license PD-1 blockade via CD28 costimulation.

The mechanisms regulating exhaustion of tumor-infiltrating lymphocytes (TIL) and responsiveness to PD-1 blockade remain partly unknown. In human ovarian cancer, we show that tumor-specific CD8+ TIL accumulate in tumor islets, where they engage antigen and upregulate PD-1, which restrains their functions. Intraepithelial PD-1+CD8+ TIL can be, however, polyfunctional. PD-1+ TIL indeed exhibit a continuum of exhaustion states, with variable levels of CD28 costimulation, which is provided by antigen-presenting cells (APC) in intraepithelial tumor myeloid niches. CD28 costimulation is associated with improved effector fitness of exhausted CD8+ TIL and is required for their activation upon PD-1 blockade, which also requires tumor myeloid APC. Exhausted TIL lacking proper CD28 costimulation in situ fail to respond to PD-1 blockade, and their response may be rescued by local CTLA-4 blockade and tumor APC stimulation via CD40L.

Resident Kupffer cells and neutrophils drive liver toxicity in cancer immunotherapy.

Immunotherapy is revolutionizing cancer treatment but is often restricted by toxicities. What distinguishes adverse events from concomitant antitumor reactions is poorly understood. Here, using anti-CD40 treatment in mice as a model of TH1-promoting immunotherapy, we showed that liver macrophages promoted local immune-related adverse events. Mechanistically, tissue-resident Kupffer cells mediated liver toxicity by sensing lymphocyte-derived IFN-γ and subsequently producing IL-12. Conversely, dendritic cells were dispensable for toxicity but drove tumor control. IL-12 and IFN-γ were not toxic themselves but prompted a neutrophil response that determined the severity of tissue damage. We observed activation of similar inflammatory pathways after anti-PD-1 and anti-CTLA-4 immunotherapies in mice and humans. These findings implicated macrophages and neutrophils as mediators and effectors of aberrant inflammation in TH1-promoting immunotherapy, suggesting distinct mechanisms of toxicity and antitumor immunity.

Trametinib Induces the Stabilization of a Dual GNAQ p.Gly48Leu- and FGFR4 p.Cys172Gly-Mutated Uveal Melanoma. The Role of Molecular Modelling in Personalized Oncology.

We report a case of an uveal melanoma patient with GNAQ p.Gly48Leu who responded to MEK inhibition. At the time of the molecular analysis, the pathogenicity of the mutation was unknown. A tridimensional structural analysis showed that Gαq can adopt active and inactive conformations that lead to substantial changes, involving three important switch regions. Our molecular modelling study predicted that GNAQ p.Gly48Leu introduces new favorable interactions in its active conformation, whereas little or no impact is expected in its inactive form. This strongly suggests that GNAQ p.Gly48Leu is a possible tumor-activating driver mutation, consequently triggering the MEK pathway. In addition, we also found an FGFR4 p.Cys172Gly mutation, which was predicted by molecular modelling analysis to lead to a gain of function by impacting the Ig-like domain 2 folding, which is involved in FGF binding and increases the stability of the homodimer. Based on these analyses, the patient received the MEK inhibitor trametinib with a lasting clinical benefit. This work highlights the importance of molecular modelling for personalized oncology.

The combination of stereotactic radiosurgery with immune checkpoint inhibition or targeted therapy in melanoma patients with brain metastases: a retrospective study.

BACKGROUND: Evidence pointing to a synergistic effect of stereotactic radiosurgery (SRS) with concurrent immunotherapy or targeted therapy in patients with melanoma brain metastases (BM) is increasing. We aimed to analyze the effect on overall survival (OS) of immune checkpoint inhibitors (ICI) or BRAF/MEK inhibitors initiated during the 9 weeks before or after SRS. We also evaluated the prognostic value of patients' and disease characteristics as predictors of OS in patients treated with SRS. METHODS: We identified patients with BM from cutaneous or unknown primary origin melanoma treated with SRS between 2011 and 2018. RESULTS: We included 84 patients. The median OS was 12 months (95% CI 9-20 months). The median follow-up was 30 months (95% CI 28-49). Twenty-eight patients with newly diagnosed BM initiated anti-PD-1 +/-CTLA-4 therapy (n = 18), ipilimumab monotherapy (n = 10) or BRAF+/- MEK inhibitors (n = 11), during the 9 weeks before or after SRS. Patients who received anti-PD-1 +/-CTLA-4 mAb showed an improved survival in comparison to ipilimumab monotherapy (OS 24 vs. 7.5 months; HR 0.32, 95% 0.12-0.83, p = 0.02) and BRAF +/-MEK inhibitors (OS 24 vs. 7 months, respectively; HR 0.11, 95% 0.04-0.34, p = 0.0001). This benefit remained significant when compared to the subgroup of patients treated with dual BRAF/MEK inhibition (BMi) (n = 5). In a multivariate Cox regression analysis an age > 65, synchronous BM, > 2 metastatic sites, > 4 BM, and an ECOG > 1 were correlated with poorer prognosis. A treatment with anti-PD-1+/-CTLA-4 mAbs within 9 weeks of SRS was associated with better outcomes. The presence of serum lactate dehydrogenase (LDH) levels ≥ 2xULN at BM diagnosis was associated with lower OS (HR 1.60, 95% CI 1.03-2.50; p = 0.04). CONCLUSIONS: The concurrent administration of anti-PD-1+/-CTLA-4 mAbs with SRS was associated with improved survival in melanoma patients with newly diagnosed BM. In addition to CNS tumor burden, the extension of systemic disease retains its prognostic value in patients treated with SRS. Elevated serum LDH levels are predictors of poor outcome in these patients.