AI Set to Transform the Medical Industry as New AI Model Detects Lung Cancer Tissue in Minutes

A team of engineers from the University of Cologne's Faculty of Medicine and University Hospital Cologne have introduced an AI model that can help detect

Lung cancer is one of the deadliest diseases in the world, claiming the lives of an estimated 1.3 million people in 2023. Non-small cell lung cancer (NSCLC) accounts for more than 80% of all lung cancers and is characterized by the development of malignant tumors in the lung tissue over time.

To remove the tumors before they destroy surrounding tissues, patients must undergo intensive and draining treatments, which can cost an average of over $68,000. Despite early diagnosis and treatment, the mortality rate for lung cancer remains high. Therefore, accurate diagnosis is crucial, second only to prevention.

Pathological examinations are the primary method used by oncologists to detect lung cancer tissue. During this process, healthcare professionals collect hematoxylin and eosin (H&E)-stained tissue samples. These samples are then reviewed by oncologists to identify the presence of tumor cells, which they use in conjunction with your data and genetics to tailor an effective treatment.

While the initial steps of tissue gathering have remained largely unchanged in oncology for decades, the way and means by which the data gets processed have migrated to a digital format. Digital pathology platforms have eliminated the need for researchers to be peering at cells through microscopes and instead use computer monitors.

Digitizing pathology has brought some serious benefits, including the ability to integrate software into the discovery process. Today, most experts use some form of digital lung tissue analysis to determine your state. In the coming years, artificial intelligence will replace manually operated software systems as the primary way of determining lung cancer's presence in tissues.

Artificial intelligence models can leverage the vast array of histological images and extract additional information that human reviewers can't capture. As such, there's a strong push to create more effective and accessible AI-powered pathological systems.

A recent study published in the journal Cell Reports Medicine unveils new AI algorithms and a computational pathology platform designed specifically for NSCLC diagnosis. The study demonstrates a combination of new AI foundation models and represents the largest and most relevant data set used to date. The system integrates a detailed multi-class tissue dataset that includes whole-slide images with vital details such as lung adenocarcinoma and squamous cell carcinomas. Notably, the AI model integrated +4k slides from 1,527 patients and was derived from an international cohort of lung cancer research providers.

The testing phase of the research involved comparing tissue sample results with expert pathologist opinions to ensure quality. The team was keen on using only explainable, independent, capable prognostic parameters derived from H&E-stained tissue samples, which made it easier to confirm results. Four AI models were used in the experiment. Each AI algorithm was designed to examine and determine different classes including epithelial tumor component, tumoral stroma, necrotic debris, and mucin. The AI system reviewed the live data and compared it to tertiary lymphoid structure and necrosis assessments within the model seeking similarities.

The results revealed that the algorithm was highly accurate and faster than other methods of determining lung cancer tumors. The team demonstrated .89 accuracy, with many of the inaccuracies falling under optical issues related to pixels rather than the AI algorithm's detection capabilities. An AI-powered lung cancer detection system brings several benefits to the market. For one, these low-cost alternatives can be used in remote regions where larger, more specialized equipment and professionals are not available. As such, they could help create a more balanced and accessible treatment process.

One of the main benefits of the AI system is that it's completely automated. Tissue samples are scanned, shown, tested, and treatment recommendations are made by the system. By reducing diagnosis times, patients can lower treatment needs and costs. Another major reason why this study has many professionals excited is that it opens the door for new data collection methods to be derived.

人工智慧演算法越來越有能力確定資料集中難以看到的模式和連結。因此，該系統將能夠不斷學習從患者收集的新舊信息，從而提高其能力。多年來對這種疾病的研究已經採集了數百萬份肺癌組織樣本。一旦將這些數據輸入到更大的人工智慧模型中，該模型可以確定難以檢測的模式和相關事件，這些數據可能會為未來的預防方法解鎖一些線索。

另一個主要好處是更好的治療。該系統使醫療保健專業人員能夠在創紀錄的時間內為患者創建優化和個人化的治療方案。肺癌診斷對於預防疾病傳播和降低患者死亡率至關重要。將來，這些系統可以放置在您的家中，甚至可以租給個人或小型診所。這項舉措將為全球採用打開大門，同時減少誤診、旅行需求以及所有相關方的費用。

該計畫的研究團隊由科隆大學醫院普通病理學和病理解剖學研究所的 Yuri Tolkach 博士和 Reinhard Büttne 教授領導。該計畫是透過北萊茵-威斯特法倫州、聯邦教育部和

的資助得以實現的

The above is the detailed content of AI Set to Transform the Medical Industry as New AI Model Detects Lung Cancer Tissue in Minutes. For more information, please follow other related articles on the PHP Chinese website!