Digital quantification of PRAME for distinguishing melanoma from nevi compared to manual assessment.

AIMS: In this proof-of-concept study, we propose a new method for automated digital quantification of PRAME (PReferentially expressed Antigen of MElanoma) as a diagnostic aid to distinguish between benign and malignant melanocytic lesions. The proposed method utilizes immunohistochemical virtual double nuclear staining for PRAME and SOX10 to precisely identify the melanocytic cells of interest, which is combined with digital image analyse to quantify a PRAME-index. METHODS: Our study included 10 compound nevi, 3 halo nevi, and 10 melanomas. Tissue slides were stained with PRAME, scanned, the cover glass removed, stained with SOX10, scanned again, and finally analysed digitally. The digitally quantified PRAME-index was compared with a manual qualitative assessment by a dermatopathologist using the standard PRAME-scoring system. RESULTS: The digitally quantified PRAME-index showed a sensitivity of 70 % and a specificity of 100 % for separating melanomas from benign lesions. The manual qualitative PRAME-score showed a sensitivity of 60 % and a specificity of 100 %. Comparing the two methods using ROC-analyses, our digital quantitative method (AUC: 0.931, 95 % CI: 0.834;1.00, SD: 0.050) remains on par with the manual qualitative method (AUC: 0.877, 95 % CI: 0.725;1.00, SD: 0.078). CONCLUSION: We found our novel digital quantitative method was at least as precise at classifying lesions as benign or malignant as the current manual qualitative assessment. Our method has the advantages of being operator-independent, objective, and replicable. Furthermore, our method can easily be implemented in an already digitalized pathology department. Given the small cohort size, more studies are to be done to validate our findings.

Immunohistochemical double nuclear staining for cell-specific automated quantification of the proliferation index - A promising diagnostic aid for melanocytic lesions.

AIMS: Pathologists often use immunohistochemical staining of the proliferation marker Ki67 in their diagnostic assessment of melanocytic lesions. However, the interpretation of Ki67 can be challenging. We propose a new workflow to improve the diagnostic utility of the Ki67-index. In this workflow, Ki67 is combined with the melanocytic tumour-cell marker SOX10 in a Ki67/SOX10 double nuclear stain. The Ki67-index is then quantified automatically using digital image analysis (DIA). The aim of this study was to optimise and test three different multiplexing methods for Ki67/SOX10 double nuclear staining. METHODS: Multiplex immunofluorescence (mIF), multiplex immunohistochemistry (mIHC), and multiplexed immunohistochemical consecutive staining on single slide (MICSSS) were optimised for Ki67/SOX10 double nuclear staining. DIA applications were designed for automated quantification of the Ki67-index. The methods were tested on a pilot case-control cohort of benign and malignant melanocytic lesions (n = 23). RESULTS: Using the Ki67/SOX10 double nuclear stain, malignant melanocytic lesions could be completely distinguished from benign lesions by the Ki67-index. The Ki67-index cut-offs were 1.8% (mIF) and 1.5% (mIHC and MICSSS). The AUC of the automatically quantified Ki67-index based on double nuclear staining was 1.0 (95% CI: 1.0;1.0), whereas the AUC of conventional Ki67 single-stains was 0.87 (95% CI: 0.71;1.00). CONCLUSIONS: The novel Ki67/SOX10 double nuclear stain highly improved the diagnostic precision of Ki67 interpretation. Both mIHC and mIF were useful methods for Ki67/SOX10 double nuclear staining, whereas the MICSSS method had challenges in the current setting. The Ki67/SOX10 double nuclear stain shows potential as a valuable diagnostic aid for melanocytic lesions.

Reducing workload in malignant melanoma sentinel node examination: a national study of pathology reports from 507 melanoma patients.

AIMS: Even though extensive melanoma sentinel node (SN) pathology protocols increase metastasis detection, there is a need for balancing high detection rates with reasonable workload. A newly tested Danish protocol recommended examining nodes at six levels 150 µm apart (six-level model) and using SOX10 and Melan-A immunohistochemistry (IHC). We explored if a protocol examining 3 levels 300 µm apart (three-level model) combined with IHC would compromise metastasis detection. The study aim was to optimise the protocol to reduce workload without compromising detection rate. METHODS: 8 months after protocol implementation, we reviewed the pathology reports of SNs from 507 melanoma patients nationwide, including 117 SN-positive patients. Each report was reviewed to determine histopathological features, including detection of metastasis, exact levels with metastasis, exact levels with metastasis >1 mm in diameter and IHC results. RESULTS: The six-level model detected metastases in 23% of patients, whereas the three-level model would have detected metastases in 22% of patients. The three-level model would have missed a few small metastases (n=4), measuring <0.1 mm, 0.1 mm, 0.4 mm and 0.1 mm, respectively. The six-level model detected metastases >1 mm in 7% of patients. One of these metastases (measuring 1.1 mm) would have been detected by the three-level model, but not as >1 mm. SOX10 and Melan-A had equal sensitivity. CONCLUSIONS: Reducing the number of levels examined to three levels 300 µm apart combined with IHC does not have significant impact on metastasis detection rate, and we will therefore recommend that the future melanoma SN guideline takes this into consideration to reduce overall workload.

Examination of the Functional Relationship between PD-L1 DNA Methylation and mRNA Expression in Non-Small-Cell Lung Cancer.

Immunotherapy targeting the interaction between programmed cell death protein 1 (PD-1) and programmed death-ligand 1 (PD-L1) is a treatment option for patients with non-small-cell lung cancer (NSCLC). The expression of PD-L1 by the NSCLC cells determines treatment effectiveness, but the relationship between PD-L1 DNA methylation and expression has not been clearly described. We investigated PD-L1 DNA methylation, mRNA expression, and protein expression in NSCLC cell lines and tumor biopsies. We used clustered regularly interspaced short palindromic repeats-associated protein 9 (CRISPR-Cas9) to modify PD-L1 genetic contexts and endonuclease deficient Cas9 (dCas9) fusions with ten-eleven translocation methylcytosine dioxygenase 1 (TET1) and DNA (cytosine-5)-methyltransferase 3A (DNMT3A) to manipulate PD-L1 DNA methylation. In NSCLC cell lines, we identified specific PD-L1 CpG sites with methylation levels inversely correlated with PD-L1 mRNA expression. However, inducing PD-L1 mRNA expression with interferon-γ did not decrease the methylation level for these CpG sites, and using CRISPR-Cas9, we found that the CpG sites did not directly confer a negative regulation. dCas9-TET1 and dCas9-DNMT3A could induce PD-L1 hypo- and hyper-methylation, respectively, with the latter conferring a decrease in expression showing the functional impact of methylation. In NSCLC biopsies, the inverse correlation between the methylation and expression of PD-L1 was weak. We conclude that there is a regulatory link between PD-L1 DNA methylation and expression. However, since these measures are weakly associated, this study highlights the need for further research before PD-L1 DNA methylation can be implemented as a biomarker and drug target for measures to improve the effectiveness of PD-1/PD-L1 immunotherapy in NSCLC.

Prognostic value of KRAS mutations, TP53 mutations and PD-L1 expression among lung adenocarcinomas treated with immunotherapy.

AIMS: The aim of this study was to investigate the association between oncogenic alterations and programmed cell death ligand 1 (PD-L1) expression in lung adenocarcinomas, as well as the prognostic value of KRAS and/or TP53 mutations in patients treated with immunotherapy. METHODS: This study is a retrospective cohort study of 519 patients with lung adenocarcinomas analysed for mutations and PD-L1 expression. Data were collected from electronic pathology record system, next-generation sequencing system, and clinical databases. Association between mutations and PD-L1 expression was investigated, as well as survival statistics of the 65 patients treated with immunotherapy. RESULTS: 41% of the samples contained a KRAS mutation, predominantly together with mutations in TP53 (41%) or STK11 (10%). Higher expression of PD-L1 was seen among patients with KRAS mutations (p=0.002) and EGFR wild type (p=0.006). For patients treated with immunotherapy, there was no statistically significant difference for overall survival (OS) and progression-free survival (PFS) according to KRAS mutation status, TP53 mutation status or PD-L1 expression. The HR for concomitant mutations in TP53 and KRAS was 0.78 (95% CI 0.62 to 0.99) for OS and 0.43 (0.21 to 0.88) for PFS. Furthermore, concomitant TP53 and KRAS mutations predicted a better PFS (p=0.015) and OS (p=0.029) compared with no mutations or a single mutation in either TP53 or KRAS. CONCLUSION: Mutations in TP53 together with KRAS may serve as a potential biomarker for survival benefits with immunotherapy.

A targeted expression panel for classification, gene fusion detection and PD-L1 measurements - Can molecular profiling replace immunohistochemistry in non-small cell lung cancer?

The histological classification of non-small-cell lung cancer (NSCLC) and identification of possible therapeutic targets are important for disease management. However, as biopsies are often small, with a limited amount of tumor cells, it can be challenging to obtain enough tissue for the needed number of diagnostic immunohistochemical stains and molecular analyses. In this study, we combined a small custom designed targeted expression panel with a commercial fusion transcript assay by which we were able to perform both a histological classification (transcribing the expression of the genes encoding TTF1, Napsin A, CK5/6, and the truncated P63 isoform ΔNp63 (p40) into either adenocarcinoma or squamous cell carcinoma) and an identification of fusion genes involving ALK, RET, and ROS1. The expression panel also included the PD-L1 encoding gene, CD274, in order to evaluate the PD-L1 mRNA potential for identification of patients who will benefit from immune checkpoint inhibitor treatment. We evaluated the panel using 42 NSCLC patient samples. The molecular profiling agreed with the original immunohistochemistry (IHC)-based classification in 93% of the cases. For ten of the patients, being fusion gene positive, the fusion transcripts were detected in 100%. The molecular assessment of PD-L1 also showed agreement with the original assessment made by IHC. In conclusion, this study presents a small, targeted expression panel with the potential to perform both a molecularly based histological classification and a fusion gene identification in NSCLC patients as well as identifying PD-L1 status from a very limited amount of starting material.

Can evaluation of mismatch repair defect and TILs increase the number of triple-negative breast cancer patients eligible for immunotherapy?

BACKGROUND: Programmed-cell-death-ligand 1 (PD-L1) inhibitor treatment is approved for metastatic/recurrent, PD-L1 positive, triple-negative breast cancer (TNBC) and solid tumors with mismatch repair (MMR) defect regardless of PD-L1 status. The analytical validity of PD-L1 has been questioned and adding evaluation of tumor-infiltrating lymphocytes (TILs) may identify patients likely to respond to immunotherapy. We investigated the association between MMR-deficiency and PD-L1 in TNBC; aiming to identify PD-L1 negative, TNBC patients that may be candidates for anti-PD-L1 immunotherapy. METHODOLOGY: Paraffin-embedded tumor material from 44 TNBC patients was included. In 38 cases, immunohistochemical-staining´s on whole-slide sections were successful for all four MMR proteins (MSH2, MSH6, MLH1 and PMS2) and PD-L1 in 36 cases. MMR-status was categorized as positive (pMMR), heterogeneous (hMMR) or deficient (dMMR). Tumor-infiltrating lymphocytes (TILs) were evaluated on H&E sections. RESULTS: MMR stainings showed substantial intratumor heterogeneity. Four of 38 cases (11%) were recorded as dMMR with loss of ≥ 1 MMR-protein and 19 cases (50%) showed heterogeneous expression or partial loss (hMMR) of ≥ 1 MMR-protein. Three of 22 PD-L1 negative cases were dMMR (14%) and ten cases hMMR (45%). 16 of 22 PD-L1 negative cases (73%) showed high TILs. CONCLUSIONS: A substantial proportion of PD-L1 negative, TNBC patients showed complete/partial loss of MMR and/or high TILs indicating that supplementing PD-L1 examination with these biomarkers may identify TNBC-patients that may be selected for immunotherapy.

Frequently used quantitative polymerase chain reaction-based methods overlook potential clinically relevant genetic alterations in epidermal growth factor receptor compared with next-generation sequencing: a retrospective clinical comparison of 1839 lung adenocarcinomas.

AIMS: The aim of the study was to investigate the advantage of implementing next-generation sequencing (NGS) compared with quantitative polymerase chain reaction (qPCR) when performing routine molecular diagnostics in adenocarcinomas of the lung. METHODS: The study is a retrospective cross-sectional observational study of 1839 cytological and histological adenocarcinoma biopsies investigated for gene mutations from 2016 to 2018 at the Department of Pathology at Aarhus University Hospital. A total of 1169 samples were analyzed by qPCR for the presence of EGFR hotspot mutations from 2016 to 2017. A total of 670 samples were analyzed with NGS for the presence of EGFR mutations and other gene mutations in 2018. RESULTS: The average frequency of EGFR mutations in the study population was 11.5%, with the highest frequency found in 2018, where NGS was implemented (10.8% in 2016, 11.5% in 2017, and 12.2% in 2018). Possible therapy resistance markers such as EGFR exon 20 mutations were found more commonly after NGS implementation, the difference being statistically significant (P = .015). In addition, NGS (2018) showed that 40.6% of the samples had KRAS mutations and 6.0% had BRAF mutations, mutations not commonly investigated in lung adenocarcinomas when qPCR is the method of choice. Among the EGFR-mutated samples analyzed with NGS, 13 contained a concurrent EGFR mutation, whereas three and two contained a concurrent KRAS and BRAF mutations, respectively. CONCLUSIONS: With the implementation in a clinical setting, NGS identifies more uncommon but potentially clinically important EGFR mutations, unique combinations of EGFR mutations, and concurrent mutations in KRAS and BRAF.

Telomerase reverse transcriptase promoter mutations and solar elastosis in cutaneous melanoma.

The aims of this study were to assess the prognostic potential of solar elastosis grading and telomerase reverse transcriptase (TERT) promoter mutations (TERTp) in melanoma and to evaluate whether an association between solar elastosis and TERTp exists. Solar elastosis in the dermis was evaluated in hematoxylin and eosin-stained whole slides from 486 malignant melanomas. Pyrosequencing was used to detect TERTp in 189 samples. There was no association between solar elastosis and TERTp (P=0.3). Severe elastosis was associated with older age (P<0.0001), ulceration (P=0.03), and location in the head/neck region (P<0.0001). The absence of elastosis was associated with younger age (P<0.0001), benign nevus remnants (P=0.001), and a positive BRAF V600E expression (P<0.0001). Severe elastosis predicted a worse relapse-free survival (hazard ratio: 2.18; 95% confidence interval: 1.30-3.64; P=0.003). However, it was not independent of age. TERTp was not associated with any adverse prognostic or clinicopathological outcome, nor any mitogen-activated protein kinase-related protein expressions. However, at a cutoff corresponding to the sensitivity of Sanger sequencing, TERTp predicted melanoma-specific death independently of age, and was associated with Breslow thickness, ulceration, tumor stage at diagnosis, BRAF V600E oncoprotein, and absence of p16 expression. In conclusion, TERTp were not related to severe elastosis and may thus be triggered by both chronic and acute intermittent sun exposure, the latter not visible on ordinary hematoxylin and eosin-stained slides. Neither TERTp nor severe elastosis predicted an adverse outcome in melanoma. An absence of elastosis was seen in younger melanoma patients and may be used to select those melanomas originating in a nevus, which often harbors a BRAF mutation.

Concordance in BRAF V600E status over time in malignant melanoma and corresponding metastases.

AIMS: The aims of the present study were to analyse the usability of an immunohistochemical (IHC) analysis as compared with a frequently used mutation detection analysis, and to examine the extent of intratumour and intertumour heterogeneity of BRAF V600E in primary tumours and their corresponding metastases. In the development of intertumour heterogeneity between the primary tumour and the corresponding metastases, time as a factor was also investigated. METHODS AND RESULTS: In total, 227 samples from 224 melanoma patients were analysed with both the Cobas 4800 BRAF V600 Mutation Test and IHC anti-BRAF V600E staining. In 82 primary tumours and 224 corresponding metastases, the extents of intertumour and intratumour heterogeneity were investigated with IHC staining. In 15 cases, disagreement between IHC analysis and the Cobas test was seen. In all but one of the examined patients, homogeneity between the primary tumour and the corresponding metastasis was found. Except for this one case, no heterogeneity developed over longer periods. CONCLUSION: IHC analysis can be safely used as a BRAF pretreatment screening tool, and no additional test is needed when staining is positive. However, if stains are negative, additional tests are essential for detection of other BRAF mutations. We suggest that using primary melanoma tissues is just as safe as using metastatic tissue for detection of BRAF V600E, as BRAF intertumour heterogeneity is extremely rare. In addition, the time between diagnosis of the primary tumour and diagnosis of the corresponding metastasis seems not to increase the risk of intertumour heterogeneity.

BRAF mutations in conjunctival melanoma: investigation of incidence, clinicopathological features, prognosis and paired premalignant lesions.

PURPOSE: To investigate incidence, clinicopathological features and prognosis of BRAF-mutated conjunctival melanoma in Denmark. Furthermore, to determine BRAF mutations in paired premalignant lesions and evaluate immunohistochemical BRAF V600E oncoprotein detection. METHODS: Data from 139 patients with conjunctival melanoma (1960-2012) were collected. Archived conjunctival melanoma samples and premalignant lesions were analysed for BRAF mutations using droplet digital polymerase chain reaction (PCR). Results were associated with clinicopathological features and compared with BRAF V600E oncoprotein stainings. RESULTS: The overall incidence of conjunctival melanoma (0.5 cases/1 000 000/year) increased during the study period with 0.13 cases/1 000 000/10 years. The increase comprised a higher proportion of patients aged >65 years, epibulbar tumours and tumours developed from a primary acquired melanosis with atypia. BRAF mutations were identified in 39 of 111 (35%) cases. The rate ratio of BRAF-mutated versus BRAF-wild-type melanoma did not change over time. BRAF mutations were associated with T1 stage (p = 0.007), young age (p = 0.001), male gender (p = 0.02), sun-exposed location (p = 0.01), mixed/non-pigmented tumour colour (p = 0.02) and nevus origin (p = 0.005), but did not associate with prognosis. BRAF status in conjunctival melanoma and paired premalignant lesions corresponded in 19 of 20 cases. Immunohistochemistry detected BRAF V600E mutations with a sensitivity of 0.94 and a specificity of 1.00 in newer conjunctival melanoma samples (2000-2012, n = 47). CONCLUSION: The incidence of conjunctival melanoma increased in Denmark over 50 years. The proportion of BRAF-mutated conjunctival melanoma was constant. BRAF mutations were identified as early events in conjunctival melanoma, associated with a distinct clinicopathological profile, similar to BRAF-mutated cutaneous melanoma. Immunohistochemical detection of BRAF can be used to assess BRAF V600E mutations.

KIT is a frequent target for epigenetic silencing in cutaneous melanoma.

The receptor tyrosine kinase KIT and its ligand, stem cell factor (SCF), are essential for the proliferation and survival of normal melanocytes. In melanomas arising on mucosal, acral, and chronically sun-damaged skin, activating KIT mutations have been identified as oncogenic drivers and potent therapeutic targets. Through an initial whole-genome screen for aberrant promoter methylation in melanoma, we identified the KIT promoter as a target for hypermethylation in 43/110 melanoma cell lines, and in 3/12 primary and 11/29 metastatic cutaneous melanomas. Methylation density at the KIT promoter correlated inversely with promoter activity in vitro and in vivo, and the expression of KIT was restored after treatment with the demethylating agent 5-aza-2'-deoxycytidine. Hypermethylation of KIT showed no direct or inverse correlations with well-documented melanoma drivers. Growth of melanoma cells in the presence of SCF led to reduced KIT expression and increased methylation density at the KIT promoter, suggesting that SCF may exert a selection pressure for the loss of KIT. The frequent loss of KIT in cutaneous melanoma by promoter hypermethylation suggests that distinct KIT signaling pathways have opposing roles in the pathogenesis of melanoma subtypes.

Immunohistochemical analysis of molecular drivers in melanoma identifies p16 as an independent prognostic biomarker.

AIMS: To perform immunohistochemical (IHC) analysis of molecular drivers related to the development and maintenance of melanoma and to assess their value as diagnostic and prognostic melanoma biomarkers in routine clinical practice. METHODS: Tissue microarrays constructed from a cohort of primary melanomas (n=355), benign naevi (n=37) and melanoma metastases (n=14) were evaluated for IHC expression of c-KIT, BRAF(V600E), MITF, p16, p53 and PTEN, as well as for pERK, a surrogate marker for mitogen-activated protein kinase pathway activation. The results were correlated with clinicopathological parameters and clinical outcome. RESULTS: Absent p16 expression and reduced MITF expression were both associated with the adverse prognostic markers ulceration (p=0.009 and p<0.0001, respectively), advanced tumour stage (p<0.0001 and p=0.001, respectively) and higher Breslow thickness (both p<0.0001), as well as with an adverse overall relapse-free survival (p<0.0001 and p=0.003, respectively). Absence of p16 expression predicted overall relapse-free (p=0.02) and distant metastasis-free (p=0.04) survival, independently of Breslow thickness, ulceration and tumour stage. CONCLUSIONS: IHC determined p16 expression is an independent prognostic biomarker of potential value in routine melanoma diagnostic practice.

TGF-ß1 and TNF-α differentially regulate Twist1 mediated resistance towards BRAF/MEK inhibition in melanoma.

Resistance to BRAF and MEK inhibition is a common phenomenon in melanoma. Cytokines and transcription factors have been attributed to contribute to the loss of sensitivity towards these inhibitors. Here, we show that transforming growth factor (TGF)-ß1 if combined with PLX4032, a BRAF inhibitor, or GSK1120212, a MEK inhibitor, substantially increased cell death in BRAF-mutant melanoma cell lines. This increase was based on the combined regulatory decrease in Twist1, an antiapoptotic protein. Overexpression or silencing of Twist1 attenuated or aggravated induction of apoptosis through PLX4032 or GSK1120212, respectively. Exposure to tumour necrosis factor (TNF)-α, however, led to increased Twist1 levels and oppositional decrease in cell death if exposed to PLX4032 or GSK1120212. This increase in drug resistance again depended on Twist1 levels. Our studies suggest that Twist1 as a common downstream target of multiple signalling cascades plays a crucial role in mediating drug resistance to BRAF- and MEK-targeted molecular inhibitors.

Evaluation of BRAF mutation testing methodologies in formalin-fixed, paraffin-embedded cutaneous melanomas.

Patients diagnosed with BRAF V600E mutated cutaneous melanoma show response to treatment with the BRAF inhibitor Vemurafenib. Different methods for BRAF mutation detection exist; however, only the Cobas 4800 BRAF V600 Mutation Test has been approved by the US Food and Drug Administration for patient selection. The results from this test depend on the percentage of tumor cells in the samples, which clinically may be estimated with substantial variation. We have evaluated five different methods: the Cobas test, Sanger sequencing, pyrosequencing, TaqMan-based allele-specific PCR, and Competitive Amplification of Differentially Melting Amplicons (CADMA), for detection of BRAF c.1799T>A (V600E) mutations in 28 formalin-fixed paraffin-embedded (FFPE) cutaneous melanoma samples. We show that the frequency of the BRAF V600E mutation is influenced by the analytical sensitivity of the applied method. However, a 100% consensus was observed among all five methods when the tumor tissue fraction was more than 10% of all tissue or more than 50% of cell-dense tissue. When using Sanger sequencing, pyrosequencing, or the Cobas test, it may be advisable to perform macrodissection before mutation testing if the tumor cell fraction is low. CADMA and TaqMan may not require macrodissections for a reliable test. Therefore, the use of more sensitive methods may have a future in testing for BRAF mutations in clinical settings.