In cancer, a biomarker refers to a substance or process that is indicative of the presence of cancer in the body. During the past decade, there has been a fundamental shift in cancer research and clinical decision-making, moving from qualitative data to quantitative digital data. An unthinkable wealth of cancer biomarkers, as well as other cancer data, have been generated by research laboratories and clinical cancer institutions worldwide. For example, between 2014 and 2018, an estimated 2 exabytes of cancer data-from genomics to diagnostic imaging-were generated in the United States. The major bulk of information arises from genomics, proteomics, metabolomics as well, as other omics, but also from oncology clinics, biochemistry, epidemiology and more. Artificial Intelligence (AI), with Machine Learning (ML) in particular and High-Performance Computing (HPC) are unique technologies able to combine all the above and particularly suited to the establishment of novel therapies and predictive models of drug response.
The idea of "one-molecule (or process) marker" indicated by its presence and the existence of an undergoing transforming cancer process is nowadays considered an utopia. Indeed, the combination of several biomarkers altogether, by means of AI and ML algorithms, are able to compile large amounts of cancer data, and appropriately trained in large cohorts of cancer patient samples, would reach unprecedented conclusions in diagnosis, prediction and general decision making of novel anticancer therapies.
In this Research Topic, we envision to gather articles of investigators working in the field of classical cancer biomarker identification (by using all the omics as well as imaging, epidemiology, clinical data etc.) but who are progressively moving towards the use of the AI and ML tools in order to reach more significant conclusions in oncological diagnosis and treatment.
Any research involving the analysis of large amounts of cancer biomarkers (any type) by using HPC, AI and/or ML, which will reach a better or improved conclusion in terms of prediction of clinical outcome when compared to similar analyzes without using AI technologies are welcome. We are interested in manuscripts from classical cancer investigators using all the omics as well as all the other techniques to identify new cancer biomarkers but who have evidenced the necessity of using AI algorithms to achieve better conclusions in terms of decision making in any kind of tumor.
Specific themes we would like to consider for publication include:
1. Latest developments of AI and ML in cancer biomarker use for diagnostic decision making and innovative cancer therapies;
2. Translating biomarker cancer research advances in AI into clinical practice;
3. Using combined digital pathological markers to improve cancer assessment;
4. Enhancement of clinical outcome prediction by means of combining genomic and proteomic cancer markers using ML technology;
5. New insights of oncological biomarker use in digital twins;
6. New approaches in the use of cancer biomarkers from multidisciplinary research teams including basic scientists, translational cancer researchers, bioinformaticians, and clinical researchers.
Manuscripts consisting solely of bioinformatics, computational analysis, or predictions of public databases which are not accompanied by validation (independent cohort or biological validation in vitro or in vivo) will not be accepted in Frontiers in Oncology.
In cancer, a biomarker refers to a substance or process that is indicative of the presence of cancer in the body. During the past decade, there has been a fundamental shift in cancer research and clinical decision-making, moving from qualitative data to quantitative digital data. An unthinkable wealth of cancer biomarkers, as well as other cancer data, have been generated by research laboratories and clinical cancer institutions worldwide. For example, between 2014 and 2018, an estimated 2 exabytes of cancer data-from genomics to diagnostic imaging-were generated in the United States. The major bulk of information arises from genomics, proteomics, metabolomics as well, as other omics, but also from oncology clinics, biochemistry, epidemiology and more. Artificial Intelligence (AI), with Machine Learning (ML) in particular and High-Performance Computing (HPC) are unique technologies able to combine all the above and particularly suited to the establishment of novel therapies and predictive models of drug response.
The idea of "one-molecule (or process) marker" indicated by its presence and the existence of an undergoing transforming cancer process is nowadays considered an utopia. Indeed, the combination of several biomarkers altogether, by means of AI and ML algorithms, are able to compile large amounts of cancer data, and appropriately trained in large cohorts of cancer patient samples, would reach unprecedented conclusions in diagnosis, prediction and general decision making of novel anticancer therapies.
In this Research Topic, we envision to gather articles of investigators working in the field of classical cancer biomarker identification (by using all the omics as well as imaging, epidemiology, clinical data etc.) but who are progressively moving towards the use of the AI and ML tools in order to reach more significant conclusions in oncological diagnosis and treatment.
Any research involving the analysis of large amounts of cancer biomarkers (any type) by using HPC, AI and/or ML, which will reach a better or improved conclusion in terms of prediction of clinical outcome when compared to similar analyzes without using AI technologies are welcome. We are interested in manuscripts from classical cancer investigators using all the omics as well as all the other techniques to identify new cancer biomarkers but who have evidenced the necessity of using AI algorithms to achieve better conclusions in terms of decision making in any kind of tumor.
Specific themes we would like to consider for publication include:
1. Latest developments of AI and ML in cancer biomarker use for diagnostic decision making and innovative cancer therapies;
2. Translating biomarker cancer research advances in AI into clinical practice;
3. Using combined digital pathological markers to improve cancer assessment;
4. Enhancement of clinical outcome prediction by means of combining genomic and proteomic cancer markers using ML technology;
5. New insights of oncological biomarker use in digital twins;
6. New approaches in the use of cancer biomarkers from multidisciplinary research teams including basic scientists, translational cancer researchers, bioinformaticians, and clinical researchers.
Manuscripts consisting solely of bioinformatics, computational analysis, or predictions of public databases which are not accompanied by validation (independent cohort or biological validation in vitro or in vivo) will not be accepted in Frontiers in Oncology.