Queiroz Fábio Ribeiro, Braga Letícia da Conceição, Melo Carolina Pereira de Souza, Gomes Matheus de Souza, Amaral Laurence Rodrigues do, Salles Paulo Guilherme de Oliveira
Instituto Mário Penna, Belo Horizonte, MG, Brazil.
Laboratório de Bioinformática e Análises Moleculares, Universidade Federal de Uberlândia, Patos de Minas, MG, Brazil.
Einstein (Sao Paulo). 2024 Sep 30;22:eAO0508. doi: 10.31744/einstein_journal/2024AO0508. eCollection 2024.
Queiroz et al. showed that the application of cluster methodology for classifying gastric cancer is suitable and efficient within a Brazilian cohort, which is known for its population heterogeneity. The study highlighted the potential utilization of this method within public health services due to its low-cost, presenting a viable means to improve the diagnosis and prognosis of gastric cancer.
Our Brazilian cohort with gastric cancer has a distinct distribution between mutated and normal p53.
New genetic marker-based classifications improve gastric cancer diagnosis accuracy.
Machine learning integration enhances predictive value in gastric cancer diagnosis.
Molecular biomarkers complement clinical decisions, advancing personalized medicine.
Gastric adenocarcinoma remains an aggressive disease with a poor prognosis, as evidenced by a 5-year survival rate of approximately 31%. The histological classifications already proposed do not accurately reflect the high biological heterogeneity of this neoplasm, particularly in diverse populations, and new classification systems using genetic markers have recently been proposed. Following these newly proposed models, we aimed to assess the cluster distribution in a Brazilian cohort. Furthermore, we evaluated whether the inclusion of other clinical and histological parameters could enhance the predictive value.
We used a previously described four-immunohistochemistry/EBER-ISH marker to classify a cohort of 30 Brazilian patients with gastric adenocarcinoma into five different clusters and compared the distribution with other genetically diverse populations. Furthermore, we used artificial intelligence methods to evaluate whether other clinical and pathological parameters could improve the results of the methodology.
Disclosing the genetic variability between populations, we observed a more balanced distribution of the aberrant/normal p53 ratio (0.6) between patients negative for the other markers tested, unlike previous studies with Asian and North American populations. In addition, decision tree analysis reinforced the efficiency of these new classifications, as the stratification accuracy was not altered with or without additional data.
Our study underscores the importance of local research in characterizing diverse populations and highlights the complementary role of molecular biomarkers in personalized medicine for gastric adenocarcinoma, enhancing diagnostic accuracy and potentially improving survival rates.
奎罗斯等人表明，聚类方法在巴西队列中应用于胃癌分类是合适且有效的，巴西队列以其人群异质性而闻名。该研究强调了由于其低成本，这种方法在公共卫生服务中的潜在用途，为改善胃癌的诊断和预后提供了一种可行的手段。
我们的巴西胃癌队列在突变型和正常型p53之间有明显的分布差异。
基于新遗传标记的分类提高了胃癌诊断的准确性。
机器学习整合增强了胃癌诊断中的预测价值。
分子生物标志物补充临床决策，推动个性化医疗。
胃腺癌仍然是一种侵袭性疾病，预后较差，5年生存率约为31%即证明了这一点。已提出的组织学分类不能准确反映这种肿瘤的高度生物学异质性，特别是在不同人群中，最近有人提出了使用遗传标记的新分类系统。遵循这些新提出的模型，我们旨在评估巴西队列中的聚类分布。此外，我们评估了纳入其他临床和组织学参数是否可以提高预测价值。
我们使用先前描述的四种免疫组织化学/EBER-ISH标记将30例巴西胃腺癌患者队列分为五个不同的聚类，并将分布与其他基因不同的人群进行比较。此外，我们使用人工智能方法评估其他临床和病理参数是否可以改善该方法的结果。
揭示人群之间的遗传变异性，我们观察到在其他测试标记为阴性的患者中，异常/正常p53比率（0.6）的分布更为平衡，这与之前对亚洲和北美人群的研究不同。此外，决策树分析加强了这些新分类的效率，因为无论有无额外数据，分层准确性都没有改变。
我们的研究强调了本地研究在表征不同人群方面的重要性，并突出了分子生物标志物在胃腺癌个性化医疗中的补充作用，提高了诊断准确性并可能提高生存率。
