Nduma Basil N, Nkeonye Stephen, Uwawah Tesingin D, Kaur Davinder, Ekhator Chukwuyem, Ambe Solomon
Internal Medicine, Medical City, Denton, USA.
Oncology, University of Texas MD Anderson Cancer Center, Houston, USA.
Cureus. 2024 Jan 26;16(1):e53024. doi: 10.7759/cureus.53024. eCollection 2024 Jan.
Colorectal cancer (CRC) is one of the most common forms of cancer. Therefore, diagnosing the condition early and accurately is critical for improved patient outcomes and effective treatment. Recently, artificial intelligence (AI) algorithms such as support vector machine (SVM) and convolutional neural network (CNN) have demonstrated promise in medical image analysis. This paper, conducted from a systematic review perspective, aimed to determine the effectiveness of AI integration in CRC diagnosis, emphasizing accuracy, sensitivity, and specificity. From a methodological perspective, articles that were included were those that had been conducted in the past decade. Also, the articles needed to have been documented in English, with databases such as Embase, PubMed, and Google Scholar used to obtain relevant research studies. Similarly, keywords were used to arrive at relevant articles. These keywords included AI, CRC, specificity, sensitivity, accuracy, efficacy, effectiveness, disease diagnosis, screening, machine learning, area under the curve (AUC), and deep learning. From the results, most scholarly studies contend that AI is superior in medical image analysis, the development of subtle patterns, and decision support. However, while deploying these algorithms, a key theme is that the collaboration between medical experts and AI systems needs to be seamless. In addition, the AI algorithms ought to be refined continuously in the current world of big data and ensure that they undergo rigorous validation to provide more informed decision-making for or against adopting those AI tools in clinical settings. In conclusion, therefore, balancing between human expertise and technological innovation is likely to pave the way for the realization of AI's full potential concerning its promising role in improving CRC diagnosis, upon which there might be significant patient outcome improvements, disease detection, and the achievement of a more effective healthcare system.
结直肠癌（CRC）是最常见的癌症形式之一。因此，早期准确诊断该病对于改善患者预后和有效治疗至关重要。最近，诸如支持向量机（SVM）和卷积神经网络（CNN）等人工智能（AI）算法在医学图像分析中展现出了前景。本文从系统综述的角度进行研究，旨在确定AI整合在CRC诊断中的有效性，重点关注准确性、敏感性和特异性。从方法学角度来看，纳入的文章是过去十年间开展的研究。此外，文章需以英文记录，利用诸如Embase、PubMed和谷歌学术等数据库获取相关研究。同样，使用关键词来检索相关文章。这些关键词包括AI、CRC、特异性、敏感性、准确性、疗效、有效性、疾病诊断、筛查、机器学习、曲线下面积（AUC）和深度学习。从结果来看，大多数学术研究认为AI在医学图像分析、细微模式识别和决策支持方面更具优势。然而，在部署这些算法时，一个关键主题是医学专家与AI系统之间需要无缝协作。此外，在当前的大数据时代，AI算法应不断优化，并确保其经过严格验证，以便在临床环境中就是否采用这些AI工具提供更明智的决策依据。因此，总之，在人类专业知识与技术创新之间取得平衡，可能为充分发挥AI在改善CRC诊断方面的潜在作用铺平道路，这可能会显著改善患者预后、实现疾病检测，并建立更有效的医疗保健系统。
