Integrating inflammation, nutrition, and artificial intelligence: A new era in colorectal cancer prognostic scoring

Abstract

Traditional tumor-node-metastasis staging overlooks key prognostic factors such as inflammation and nutrition, limiting individualized treatment in colorectal cancer. Integrating biochemical markers with artificial intelligence can significantly improve survival prediction and treatment personalization by analyzing complex, multimodal data. This evolving approach holds transformative potential for precision oncology.

Key Words: Colorectal cancer; Prognostic nutritional scoring system; Artificial intelligence; Global leadership initiative on malnutrition; Nutritional biomarkers; Tumor-infiltrating lymphocytes; Personalized cancer care

Core Tip: Integrating biomarkers of inflammation and nutrition with artificial intelligence offers a significant advancement in prognostic scoring for colorectal cancer. This approach supports truly personalized treatment strategies by analyzing complex patient data beyond traditional staging and enables more accurate survival prediction. Combining artificial intelligence-driven analysis with a full biomarker profile is a game-changing move towards precision oncology. This method enhances clinical decision-making and paves the way for improved outcomes in the management of colorectal cancer.

TO THE EDITOR

Li et al[1] recently published a study that introduces a new way to score the prognoses of patients with colorectal cancer (CRC). This method combines biomarkers related to inflammation and nutrition to help patients in stages I through III of the disease. The importance of inflammation and nutritional status in cancer prognosis is well known in the literature[2-4]. However, researchers have not sufficiently investigated the systematic incorporation of these two factors into a prognostic scoring system for CRC to date. Li et al[1] took an important step in improving our understanding of CRC prognosis by including markers such as C-reactive protein (CRP), albumin, and other biomarkers of inflammation and nutritional status. Previous studies have shown that high CRP levels are associated with poor outcomes in patients with CRC[5,6], and malnutrition, which may compromise treatment tolerance, has been considered an adverse prognostic factor[7]. As a next step, integrating artificial intelligence (AI) may further refine this scoring system by leveraging its capacity to process multimodal data for individualized prediction.

Numerous prognostic scoring methods for CRC have been developed throughout the years, aiming to enhance survival prediction accuracy and inform clinical decision-making. Prominent models encompass the tumor-node-metastasis staging system, which is the benchmark for staging CRC, with more contemporary and intricate models such as the OncoScore and National Comprehensive Cancer Network guidelines, which strive to integrate molecular and genetic elements into prognostication[8,9]. However, traditional prognostic models such as the tumor-node-metastasis system, while crucial, offer limited insight into individualized patient risk. The addition of inflammation-related markers like CRP and nutritional status indicators, such as the Global Leadership Initiative on Malnutrition criteria, has been a recent trend in CRC prognostic research[10]. Prior research indicates that systemic inflammation, especially increased CRP levels, correlates with adverse outcomes in patients with CRC[11]. Malnutrition is also acknowledged as a detrimental prognostic factor, as inadequate nutritional status might result in decreased survival rates[12].

Li et al[1] differ in that it offers a more comprehensive scoring model by integrating multiple inflammatory and nutritional biomarkers, making it a potentially more robust tool compared to previous systems. Their approach also takes into account the dynamic interaction between inflammation and nutrition, which has been suggested by several studies to play a synergistic role in cancer progression and treatment response[13,14]. Furthermore, recent literature has progressively endorsed the therapeutic efficacy of prognostic models that include both systemic inflammation and nutritional status. Portale et al[15] similarly discovered that inflammatory markers, especially the neutrophil-to-lymphocyte ratio, are substantially associated with CRC prognosis, indicating that integrating these parameters with conventional clinical staging may enhance the accuracy of survival predictions.

Li et al’s study[1] enhances the existing evidence supporting a multimodal approach to CRC prognosis, highlighting the importance of a comprehensive perspective that incorporates inflammation and nutrition for precise outcome prediction. Their model facilitates more investigation into the significance of other biomarkers, including tumor-infiltrating lymphocytes, which are gaining recognition for their prognostic importance in CRC[16]. Furthermore, research has substantiated the prognostic significance of inflammation and nutritional condition for survival outcomes in CRC and inflammatory bowel disease patients, corroborating the conclusions about the critical influence of these parameters on the prognosis[17-19].

AI, particularly machine learning and deep learning algorithms, has demonstrated increasing efficacy in CRC prognostication by integrating heterogeneous data sources such as clinical parameters, molecular profiles, imaging, and biomarker panels[20,21]. Recent studies have shown that AI models can outperform traditional statistical methods in predicting recurrence risk and treatment response, enabling dynamic, patient-specific risk stratification[22,23]. The integration of AI may provide a more sophisticated amalgamation of inflammatory and nutritional biomarkers, together with additional patient-specific variables, to enhance risk categorization. AI systems, especially machine learning models, can analyze extensive datasets, uncovering concealed patterns and connections that traditional statistical methods can overlook. Within the realm of CRC, machine learning models have been effective in forecasting recurrence and treatment response, frequently surpassing the accuracy of conventional methods[24]. Integrating inflammation-related biomarkers and dietary indicators with AI-driven analysis could significantly improve the prognostic score system, providing highly tailored and dynamic risk assessments. This may assist clinicians in predicting survival with more accuracy and in formulating the most effective treatment strategies for specific patients. For example, AI tools such as deep learning can be used to analyze imaging data such as computed tomography scans or magnetic resonance images together with biomarker data, thus predicting the likelihood of recurrence or metastasis with unprecedented precision[25]. Furthermore, AI can be used to track changes in these biomarkers over time, allowing for real-time updates to the prognostic model as patient conditions evolve. The incorporation of AI into clinical practice is an ongoing research domain, although its capacity to enhance prognostic models in CRC is unequivocal. The integration of AI with inflammation- and nutrition-related biomarkers may facilitate the advancement of really personalized medicine strategies, where in treatment decisions are informed by a comprehensive and continuously updated understanding of each patient’s distinct clinical and biological profile[26].

In conclusion, scoring methods that integrate inflammation and nutritional biomarkers may provide a viable tool for improved survival prediction, increasing the possibility of personalized treatment regimens. Incorporating AI into this model could increase prognostic accuracy and facilitate personalized treatment, moving us toward a future where precision oncology is the norm. The future of CRC management is expected to be significantly impacted by the integration of AI with biomarker-driven models as research in this area continues to evolve.