Table 2 Summary of the role of AI in refining mutanome-based immunotherapy of malignant melanoma

Enhance understanding of melanoma mutanome [120,121,122,123,124, 129, 132, 133, 135,136,137]

• Deep finding algorithms can promote mutanome examination, that is used to advance rapid technology-based identification and validation of individual mutanomes by IVAC

• Neural networks recognise mutation-derived neoantigens by incorporating genomic, epigenomic, and immunogenicity information at an unprecedented scale and swiftness

• Uncover mutations related to immunotherapy response or resistance, guiding individual categorisation and combination tactics

• Potential to reveal new pathways to conquer resistance by targeting special mutational signatures

• Supply insights into optimising mutanome-focused methods through rational drug combinations impacting ribosome biogenesis or epigenetics

• Radiomics extract patterns from imaging modalities like CT, MRI, and PET

• Patterns derived from radiomics serve as a basis for response rate monitoring, risk stratification, survival analysis, metastatic capability predictions, and patient monitoring

• In individualised therapy, radiomics discerns subtle differences in images, forming patterns influencing therapy choices

Facilitate the development of melanoma vaccines [95,96,97, 125, 128]

• Personalised melanoma vaccines

• Radiomics contributes to individualised therapy by generating predictive signatures

• Optimised identification of neoantigens, leading to the development of individualised vaccines for different mutational variants

• Streamline vaccines suited to individual immune profiles

• Hasten the development of large amounts of vaccines for individuals in a short period of time

Refining adoptive cell therapy immunotherapy option [9, 70, 89, 125,126,127, 141, 142]

• Refine ACT

• Enhance the modulation of T cells, having greater specificity for individual mutations

• Limit resource waste and identify major lapses and potential adverse effects early through simulation

• Mitigate T cell specificity loss, optimised by TCR deep sequencing

• Development of novel ACTs that recognise individual neoantigens, enabled by advancements in predictive algorithms for minigenes to analyse T cell reactivity in tumours

• Application in signature-immune marker correlations extends to other cancers like non-small-cell lung and renal cancers

Determine prediction capacities [123, 141, 143, 145]

• Improve prediction capacity, thus increasing drug discovery pipeline efficiency

• Predict specific type of mutations that initiate cancer in an individual via a noninvasive method (machine learning-assisted radiomics technique)

• Improve predictions in patients at higher risk of metastasis based on their mutanome

• Streamline metastatic risk assessment

• Allows for early preventive measures that can increase patient survival rates

• Automate the identification and segmentation of lesions in melanoma

• Radiomics serve as a predictive signature generator, aiding in better correlation with immune markers

• Signature correlations have been utilised in evaluating survival in melanoma patients treated with pembrolizumab