Machine learning models demonstrate that clinicopathologic variables are comparable to gene expression prognostic signature in predicting survival in uveal melanoma

Donizy, Piotr; Mateusz, Krzyziński,; Markiewicz, Anna; Karpiński, Paweł; Kotowski, Krzysztof; Kowalik, Artur; Orłowska-Heitzman, Jolanta; Romanowska–Dixon, Bożena; Biecek, Przemysław; Hoang, Mai P.

doi:10.1016/j.ejca.2022.07.031

Cited by 11 publications

(9 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the ratio of exhausted CD8 T-cells to cytotoxic T cells to total CD8 T-cells and Th1 cells is significantly higher in UM metastases [55]. In our recent study, we showed that TIL density and increased density of tumor-associated macrophages in the primary tumor is a significant predictor of PFS in UM patients [6]. Accordingly, the composition of the lymphoid infiltrate is of great importance rather than its mere presence.…”

Section: Discussionmentioning

confidence: 99%

“…Risk estimation of metastatic spread can be obtained through the assessment of several parameters including clinical and histopathologic features, chromosome copy number alterations, mutational status of known UM driver genes, and gene expression profiles [3][4][5]. Clinically and histopathologically, age, the greater tumor thickness, larger basal diameter, higher mitotic rate, prominent nucleoli, infiltration of the ciliary body, and loss of BAP1 expression, greater infiltration of lymphocytes and macrophages correlated with a higher probability of developing metastatic disease [1,6]. Chromosomal status with the occurrence of monosomy 3, gain of 8q, and loss of 1p are associated with an increased risk of metastasis and a poor prognosis; whereas gain of 6p is associated with good prognosis [5].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Molecular profiling of primary uveal melanoma: results of a Polish cohort

et al. 2023

Self Cite

View full text Add to dashboard Cite

There is no published data regarding the molecular alterations of Polish patients with primary uveal melanoma. We performed whole exome sequencing of 20 primary uveal melanomas (UMs), 10 metastasizing and 10 non-metastasizing cases to identify significant molecular alterations. We detected mutations and copy number variants in the BAP1 gene in 50% (10 cases) of the cases. GNA11 mutations were detected in 50% (10 cases) including nine p.Q209L and one p.R183C. GNAQ mutations gene were detected in 40% (8 cases) and all were p.Q209P. SF3B1, EIF1AX, PLCB4, and PALB2 mutations were detected in one case each. Genetic aberrations of FBXW7 were detected in 55% of cases, with copy number loss of 10 and missense mutation in one. Gain or loss of copy number was observed in 60%, 60%, and 10% of cases in MYC, MLH1, and CDKN2A genes, respectively. BAP1 and GNAQ tumor suppressor genes are more often mutated in UM with metastasis, while GNA11 mutations are more frequently detected in non-metastasizing tumors. MYC copy gain was present twice as frequently (80% versus 40%) in cases with versus those without metastases. BAP1 mutation correlated with worse overall survival; while GNA11 mutation and CDKN2A loss correlated with better and worse progression-free survival, respectively. We have confirmed BAP1 prognostic potential and documented frequent MYC amplification in metastasizing cases. Although GNA11 mutation and CDKN2A loss significantly correlated with progression-free survival in our study, our sample size is small. The prognostic significance of GNAQ/GNA11 mutation and CDKN2A loss would require further investigation. Melanoma Res 33: 104-115

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Molecular profiling of primary uveal melanoma: results of a Polish cohort

et al. 2023

Self Cite

View full text Add to dashboard Cite

show abstract

“…SKCM is one of the deadliest malignancies and prone to metastasis. Although chemotherapy, immunotherapy and molecular therapy are available, the prognosis for SKCM patients remains poor, with a very short median survival time ( Donizy et al, 2022 ; Voglis et al, 2022 ; Zhang et al, 2022 ). Although there are a variety of clinical tools to predict the prognosis of patients with SKCM ( Weiss et al, 2015 ), in view of the clinical and biological heterogeneity of primary SKCM, new methods or models that more accurately predict the prognosis of patients with SKCM are still needed.…”

Section: Discussionmentioning

confidence: 99%

The value of cuproptosis-related differential genes in guiding prognosis and immune status in patients with skin cutaneous melanoma

et al. 2023

View full text Add to dashboard Cite

Background: Skin cutaneous melanoma (SKCM) is one of the most common cutaneous malignancies, which incidence is increasing. Cuproptosis is a new type of programming cell death recently reported, which may affect the progression of SKCM.Method: The mRNA expression data of melanoma were obtained from the Gene Expression Omnibus and the Cancer Genome Atlas databases. We constructed a prognostic model according to the cuproptosis-related differential genes in SKCM. Finally, real-time quantitative PCR was performed to verify the expression of cuproptosis-related differential genes in patients with different stages of cutaneous melanoma.Results: We detected 767 cuproptosis-related differential genes based on 19 cuproptosis-related genes, and screened out 7 differential genes to construct a prognostic model, which including three high-risk differential genes (SNAI2, RAP1GAP, BCHE), and four low-risk differential genes (JSRP1, HAPLN3, HHEX, ERAP2). Kaplan-Meier analysis indicated that SKCM patients with low-risk differential genes signals had better prognosis. The Encyclopedia of Genomes results manifested that cuproptosis-related differential genes are not only involved in T cell receptor signaling channel, natural killer cell mediated cytotoxicity, but also chemokine signaling pathway and B cell receptor signaling pathway. In our risk scoring model, the receiver operating characteristic (ROC) values of the three-time nodes are 0.669 (1-year), 0.669 (3-year) and 0.685 (5-year), respectively. Moreover, the tumor burden mutational and immunology function, cell stemness characteristics and drug sensitivity have significant differences between low-risk group and high-risk group. The mRNA level of SNAI2, RAP1GAP and BCHE in stage Ⅲ+Ⅳ SKCM patients was significantly higher than that in stage Ⅰ+Ⅱ patients, while the level of JSRP1, HAPLN3, HHEX and ERAP2 in stage Ⅰ+Ⅱ SKCM patients was more remarkable higher than that in stage Ⅲ+Ⅳ SKCM patients.Conclusion: In summary, we suggest that cuproptosis can not only regulate the tumor immune microenvironment but also affect the prognosis of SKCM patients, and may offer a basic theory for SKCM patients survival studies and clinical decision-making with potentially therapeutic drugs.

show abstract

“…Donizy et al [ 11 ] used classical ML models to identify predictors for metastasis and survival in patients with UM, based on routine histological and clinical measurements in cases where molecular assays were not readily available. In this study, enucleated eyes of 164 UM patients without prior treatment were included.…”

Section: Methodsmentioning

confidence: 99%

“…Using data from three visits, the model achieved an AUC of 0.85, accuracy of 79.5%, sensitivity of 77.1%, and specificity of 79.8% in metastasis prediction. Donizy et al [11] used classical ML models to identify predictors for metastasis and survival in patients with UM, based on routine histological and clinical measurements in cases where molecular assays were not readily available. In this study, enucleated eyes of 164 UM patients without prior treatment were included.…”

Section: Both Classical Machine Learning Techniques and Deep Learning...mentioning

confidence: 99%

Artificial intelligence for ocular oncology

Koseoglu

Corrêa

Liu

2023

Current Opinion in Ophthalmology

View full text Add to dashboard Cite

Purpose of review The aim of this article is to provide an update on the latest applications of deep learning (DL) and classical machine learning (ML) techniques to the detection and prognostication of intraocular and ocular surface malignancies. Recent findings Most recent studies focused on using DL and classical ML techniques for prognostication purposes in patients with uveal melanoma (UM). Summary DL has emerged as the leading ML technique for prognostication in ocular oncological conditions, particularly in UM. However, the application of DL may be limited by the relatively rarity of these conditions.

show abstract

Machine learning models demonstrate that clinicopathologic variables are comparable to gene expression prognostic signature in predicting survival in uveal melanoma

Cited by 11 publications

References 28 publications

Molecular profiling of primary uveal melanoma: results of a Polish cohort

Molecular profiling of primary uveal melanoma: results of a Polish cohort

The value of cuproptosis-related differential genes in guiding prognosis and immune status in patients with skin cutaneous melanoma

Artificial intelligence for ocular oncology

Contact Info

Product

Resources

About