An Artificial Intelligence-Based Tool for Data Analysis and Prognosis in Cancer Patients: Results from the Clarify Study.

BACKGROUND: Artificial intelligence (AI) has contributed substantially in recent years to the resolution of different biomedical problems, including cancer. However, AI tools with significant and widespread impact in oncology remain scarce. The goal of this study is to present an AI-based solution tool for cancer patients data analysis that assists clinicians in identifying the clinical factors associated with poor prognosis, relapse and survival, and to develop a prognostic model that stratifies patients by risk. MATERIALS AND METHODS: We used clinical data from 5275 patients diagnosed with non-small cell lung cancer, breast cancer, and non-Hodgkin lymphoma at Hospital Universitario Puerta de Hierro-Majadahonda. Accessible clinical parameters measured with a wearable device and quality of life questionnaires data were also collected. RESULTS: Using an AI-tool, data from 5275 cancer patients were analyzed, integrating clinical data, questionnaires data, and data collected from wearable devices. Descriptive analyses were performed in order to explore the patients' characteristics, survival probabilities were calculated, and a prognostic model identified low and high-risk profile patients. CONCLUSION: Overall, the reconstruction of the population's risk profile for the cancer-specific predictive model was achieved and proved useful in clinical practice using artificial intelligence. It has potential application in clinical settings to improve risk stratification, early detection, and surveillance management of cancer patients.

Comprehensive cross-platform comparison of methods for non-invasive EGFR mutation testing: results of the RING observational trial.

Several platforms for noninvasive EGFR testing are currently used in the clinical setting with sensitivities ranging from 30% to 100%. Prospective studies evaluating agreement and sources for discordant results remain lacking. Herein, seven methodologies including two next-generation sequencing (NGS)-based methods, three high-sensitivity PCR-based platforms, and two FDA-approved methods were compared using 72 plasma samples, from EGFR-mutant non-small-cell lung cancer (NSCLC) patients progressing on a first-line tyrosine kinase inhibitor (TKI). NGS platforms as well as high-sensitivity PCR-based methodologies showed excellent agreement for EGFR-sensitizing mutations (K = 0.80-0.89) and substantial agreement for T790M testing (K = 0.77 and 0.68, respectively). Mutant allele frequencies (MAFs) obtained by different quantitative methods showed an excellent reproducibility (intraclass correlation coefficients 0.86-0.98). Among other technical factors, discordant calls mostly occurred at mutant allele frequencies (MAFs) ≤ 0.5%. Agreement significantly improved when discarding samples with MAF ≤ 0.5%. EGFR mutations were detected at significantly lower MAFs in patients with brain metastases, suggesting that these patients risk for a false-positive result. Our results support the use of liquid biopsies for noninvasive EGFR testing and highlight the need to systematically report MAFs.

Normal tissue content impact on the GBM molecular classification.

Molecular classification of glioblastoma has enabled a deeper understanding of the disease. The four-subtype model (including Proneural, Classical, Mesenchymal and Neural) has been replaced by a model that discards the Neural subtype, found to be associated with samples with a high content of normal tissue. These samples can be misclassified preventing biological and clinical insights into the different tumor subtypes from coming to light. In this work, we present a model that tackles both the molecular classification of samples and discrimination of those with a high content of normal cells. We performed a transcriptomic in silico analysis on glioblastoma (GBM) samples (n = 810) and tested different criteria to optimize the number of genes needed for molecular classification. We used gene expression of normal brain samples (n = 555) to design an additional gene signature to detect samples with a high normal tissue content. Microdissection samples of different structures within GBM (n = 122) have been used to validate the final model. Finally, the model was tested in a cohort of 43 patients and confirmed by histology. Based on the expression of 20 genes, our model is able to discriminate samples with a high content of normal tissue and to classify the remaining ones. We have shown that taking into consideration normal cells can prevent errors in the classification and the subsequent misinterpretation of the results. Moreover, considering only samples with a low content of normal cells, we found an association between the complexity of the samples and survival for the three molecular subtypes.

PIM kinases as potential therapeutic targets in a subset of peripheral T cell lymphoma cases.

Currently, there is no efficient therapy for patients with peripheral T cell lymphoma (PTCL). The Proviral Integration site of Moloney murine leukemia virus (PIM) kinases are important mediators of cell survival. We aimed to determine the therapeutic value of PIM kinases because they are overexpressed in PTCL patients, T cell lines and primary tumoral T cells. PIM kinases were inhibited genetically (using small interfering and short hairpin RNAs) and pharmacologically (mainly with the pan-PIM inhibitor (PIMi) ETP-39010) in a panel of 8 PTCL cell lines. Effects on cell viability, apoptosis, cell cycle, key proteins and gene expression were evaluated. Individual inhibition of each of the PIM genes did not affect PTCL cell survival, partially because of a compensatory mechanism among the three PIM genes. In contrast, pharmacological inhibition of all PIM kinases strongly induced apoptosis in all PTCL cell lines, without cell cycle arrest, in part through the induction of DNA damage. Therefore, pan-PIMi synergized with Cisplatin. Importantly, pharmacological inhibition of PIM reduced primary tumoral T cell viability without affecting normal T cells ex vivo. Since anaplastic large cell lymphoma (ALK+ ALCL) cell lines were the most sensitive to the pan-PIMi, we tested the simultaneous inhibition of ALK and PIM kinases and found a strong synergistic effect in ALK+ ALCL cell lines. Our findings suggest that PIM kinase inhibition could be of therapeutic value in a subset of PTCL, especially when combined with ALK inhibitors, and might be clinically beneficial in ALK+ ALCL.

Pim 1 kinase inhibitor ETP-45299 suppresses cellular proliferation and synergizes with PI3K inhibition.

The serine/threonine Pim 1 kinase is an oncogene whose expression is deregulated in several human cancers. Overexpression of Pim 1 facilitates cell cycle progression and suppresses apoptosis. Hence pharmacologic inhibitors of Pim 1 are of therapeutic interest for cancer. ETP-45299 is a potent and selective inhibitor of Pim 1 that inhibits the phosphorylation of Bad and 4EBP1 in cells and suppresses the proliferation of several non-solid and solid human tumor cell lines. The combination of the PI3K inhibitor GDC-0941 with ETP-45299 was strongly synergistic in MV-4-11 AML cells, indicating that the combination of selective Pim kinase inhibitors and PI3K inhibitor could have clinical benefit.

Thyroid receptor ligands. Part 4: 4'-amido bioisosteric ligands selective for the thyroid hormone receptor beta.

Based on the examination of the X-ray crystallographic structures of the LBD of TRalpha and TRbeta in complex with KB-141 (2), a number of novel 4'-hydroxy bioisosteric thyromimetics were prepared. Optimal affinity and beta-selectivity (33 times), was found with a medium-sized alkyl-substituted amido group; iso-butyl (12c). It can be concluded that bioisosteric replacements of the 4'-hydroxy position represent a new promising class of TRbeta-selective synthetic thyromimetics.

Thyroid receptor ligands. 1. Agonist ligands selective for the thyroid receptor beta1.

Endogenous thyroid receptor hormones 3,5,3',5'-tetraiodo-l-thyronine (T(4), 1) and 3,5,3'-triiodo-l-thyronine (T(3), 2) exert a significant effects on growth, development, and homeostasis in mammals. They regulate important genes in intestinal, skeletal, and cardiac muscles, the liver, and the central nervous system, influence overall metabolic rate, cholesterol and triglyceride levels, and heart rate, and affect mood and overall sense of well being. The literature suggests many or most effects of thyroid hormones on the heart, in particular on the heart rate and rhythm, are mediated through the TRalpha(1) isoform, while most actions of the hormones on the liver and other tissues are mediated more through the TRbeta(1) isoform of the receptor. Some effects of thyroid hormones may be therapeutically useful in nonthyroid disorders if adverse effects can be minimized or eliminated. These potentially useful features include weight reduction for the treatment of obesity, cholesterol lowering for treating hyperlipidemia, amelioration of depression, and stimulation of bone formation in osteoporosis. Prior attempts to utilize thyroid hormones pharmacologically to treat these disorders have been limited by manifestations of hyperthyroidism and, in particular, cardiovascular toxicity. Consequently, development of thyroid hormone receptor agonists that are selective for the beta-isoform could lead to safe therapies for these common disorders while avoiding cardiotoxicity. We describe here the synthesis and evaluation of a series of novel TR ligands, which are selective for TRbeta(1) over TRalpha(1). These ligands could potentially be useful for treatment of various disorders as outlined above. From a series of homologous R(1)-substituted carboxylic acid derivatives, increasing chain length was found to have a profound effect on affinity and selectivity in a radioreceptor binding assay for the human thyroid hormone receptors alpha(1) and beta(1) (TRalpha(1) and TRbeta(2)) as well as a reporter cell assay employing CHOK1-cells (Chinese hamster ovary cells) stably transfected with hTRalpha(1) or hTRbeta(1) and an alkaline phosphatase reporter-gene downstream thyroid response element (TRAFalpha(1) and TRAFbeta(1)). Affinity increases in the order formic, acetic, and propionic acid, while beta-selectivity is highest when the R(1) position is substituted with acetic acid. Within this series 3,5-dibromo-4-[(4-hydroxy-3-isopropylphenoxy)phenyl]acetic acid (11a) and 3,5-dichloro-4-[(4-hydroxy-3-isopropylphenoxy)phenyl]acetic acid (15) were found to reveal the most promising in vitro data based on isoform selectivity and were selected for further in vivo studies. The effect of 2, 11a, and 15 in a cholesterol-fed rat model was monitored including potencies for heart rate (ED(15)), cholesterol (ED(50)), and TSH (ED(50)). Potency for tachycardia was significantly reduced for the TRbeta selective compounds 11a and 15 compared with 2, while both 11a and 15 retained the cholesterol-lowering potency of 2. This left an approximately 10-fold therapeutic window between heart rate and cholesterol, which is consistent with the action of ligands that are approximately 10-fold more selective for TRbeta(1). We also report the X-ray crystallographic structures of the ligand binding domains of TRalpha and TRbeta in complex with 15. These structures reveal that the single amino acid difference in the ligand binding pocket (Ser277 in TRalpha or Asn331 in TRbeta) results in a slightly different hydrogen bonding pattern that may explain the increased beta-selectivity of 15.