INTRODUCTION
Gastrointestinal (GI) cancers remain a major global health challenge, accounting for a significant proportion of cancer-related morbidity and mortality worldwide, which is estimated at about 35%-40% of all cancer-related deaths[1-3]. The incidence of GI cancers is assumed to continue rising due to aging populations, changes in dietary patterns, obesity, and other risk factors[4,5]. Mortality rates may decrease slightly in high-income countries due to early detection and appropriate treatments but are likely to remain high in low- and middle-income countries due to inadequate access to healthcare[6,7]. Successful policies for reducing cancer mortality in wealthy countries are not necessarily feasible or affordable in other countries[6]. These issues highlight the ongoing need for global health initiatives focused on prevention, early detection, and improved treatment strategies for GI cancers[8,9].
The landscape of oncology, particularly in gastroenterology, has undergone a remarkable transformation in recent years[10]. This transformation is driven by the convergence of technological innovation and cutting-edge clinical practices. Technological advances have permeated every aspect of cancer diagnosis, treatment and patient management, promising to reshape outcomes in ways previously deemed unimaginable. This scope of innovation comprises the virtual and augmented reality, the internet of medical things and remote patient monitoring, financial technology integration, cloud migration, and the crucial role of machine learning techniques, among others[11,12]. Additionally, it promises to redefine clinical practice and take patient care to unprecedented levels. However, it is important to provide unequivocal proof of the health benefits of emerging technological progress[12,13]. Nonetheless, the reality of integrating these advancements into everyday clinical practice remains complex and challenging. This is relevant, not only in technical aspects but also in socioeconomic terms[11].
The question before us is whether these innovations have tangibly improved oncological outcomes or if the current state of art in treatment continues to hold sway as the gold standard[14,15]. This editorial explores the dichotomy between cutting-edge technology and established clinical practices in the pursuit of optimal outcomes in oncological gastroenterology.
THE PROMISE OF TECHNOLOGICAL ADVANCEMENTS
The GI cancer treatment landscape has evolved significantly in the last decades, driven by the convergence of medical and surgical innovations, as the emergence of artificial intelligence (AI), robotics, next-generation sequencing (NGS) and personalized medicine[15-17]. These technologies offer exceptional precision in diagnosis and treatment, allowing for more tailored therapies that align with the unique genetic and molecular profiles of individual tumours.
AI AND MACHINE LEARNING
Machine learning techniques have revolutionized the GI cancers management, promising to transform diagnostics, treatment approaches, and overall patient care. Emerging as powerful tools in healthcare, this includes the ability to analyze sizable datasets, identify complex patterns and anomalies that may elude the human eye, attain an earlier and more accurate diagnosis and obtain meaningful insights to support medical decision-making[12].
Machine learning algorithms, using highly accurate automated image recognition and classification systems, help healthcare professionals to detect abnormal conditions at an early stage, facilitating prompt management. These advances hold the potential to significantly improve survival rates, particularly in cancers like colorectal cancer, where early detection is critical[18,19]. Nowadays, in low-income countries, screening programmes and policies around colorectal cancer detection are lacking. Prediction models, like the multianalyte assays with algorithmic analysis, are currently proposed as a population-based surveillance and early detection tool[20]. This system seems to offer one approach to help bridge the diagnostic-treatment gap along with more complex technological solutions. In addition, these systems might be fitted for the prediction of colorectal cancer prognosis and treatment outcomes[20-23]. Despite this, the accurate colorectal cancer diagnosis remains challenging, AI has demonstrated the ability to help in qualitative diagnosis and staging. Progress in these fields comprises the integration of machine learning algorithms in endoscopy and virtual colonoscopy, to accurately identify polyps, improve adenoma recognition rate and differentiate the neoplastic ones[24,25]. Another key issue is the preoperative staging of GI cancer. And as far as rectal cancer is concerned, machine learning techniques can scrutinize several imaging modalities, such as magnetic resonance imaging or computed tomography, to accurately assess tumour characteristics, lymph node involvement and local tumour invasion[25]. Additionally, AI can provide more detailed and consistent staging information, assisting in treatment planning and decision-making process. In rectal cancer patients, these models seem to be useful to identify high-risk cases, predict outcomes and complete response after neoadjuvant therapy, and guide personalized treatment strategies[26]. In cancer progression and prognosis, machine learning tools provide useful information on the transformation and metabolic pathways induced by drugs, improving knowledge in biological behaviour. Finally, surgical decision-making (e.g. covering ileostomy after anterior resection, intraoperative guidance, and automate skill assessment) involves a comprehensive evaluation of various specific factors to each patient’s condition[27-30].
Some specific machine learning models are crucial to understanding AI’s potential in GI cancer diagnosis, particularly in the context of endoscopic image investigation. Several models have been introduced and widely used in clinical practice, as the convolutional neural networks (CNNs), the backbone of image analysis in medical AI. They are widely used for image recognition tasks because they excel at detecting patterns, such as textures and shapes, in images. In GI endoscopy, CNNs can identify abnormal tissue patterns, detect polyps, and classify lesions[14]. This approach, combined with recurrent neural networks, is helpful to analyze video frames in sequential order, making them especially useful for spotting abnormalities over a sequence of frames, in real-time endoscopic procedures. These hybrid models are essential to improve diagnostic accuracy[31]. In addition, some researchers are exploring reinforcement learning (RL) to improve endoscopic navigation. RL agents can assist endoscopists in maneuvering the endoscope to optimize the field of view, increasing the likelihood of detecting hard-to-see lesions or polyps[32].
Despite the promising potential, there are some pressure points concerning its successful integration into clinical practice. Finance and governance, ethical issues, data privacy, algorithm complexity, quality assurance and health service delivery are some of the critical domains that should be addressed.
Reflect on the ethical considerations surrounding data privacy and governance in healthcare technology integration is indeed essential, particularly as data-driven technologies like AI become more central to patient care. First, primary ethical issues turn around who owns patient data, particularly as it becomes gathered for AI training and analytics. Patients, healthcare institutions, and AI developers often have competing claims over data ownership, which can complicate consent and governance. A review by Mittelstadt et al[33] highlights the need for transparent data ownership frameworks to protect patient rights while fostering innovation. Second, patient inform consents and transparency are crucial. Unlike traditional clinical data management, AI development often demands large datasets and, at times, data sharing across institutions or borders. In this regard, legal and ethical frameworks are necessary for obtaining informed consent, warranting they understand how their data may be used beyond their immediate care[1].
Data privacy and security are additional points that are essential in this heading. With the increase in sensitive health data being stored and analyzed, risks of data breaches and unauthorized access have risen. Robust encryption, de-identification protocols, and regular auditing are essential to maintaining data integrity and protecting patient privacy. Implementing data governance strategies that align with national regulations is also critical[4]. The utilization of AI in this field gives rise to several significant ethical considerations, particularly in relation to the decision-making process. The susceptibility of AI algorithms to biases present in training data may result in unequal treatment or misdiagnoses for underrepresented patient populations[34]. Another challenge is the “black box” nature of many algorithms, wherein decisions are made without clear human-readable explanations. Tools such as explainable AI are being developed with the objective of making AI outputs more interpretable. The introduction of transparent algorithms can enhance the level of trust placed in AI-based decisions, empowering clinicians to assess these decisions in a more critical manner[35]. The accountability and liability issues surrounding AI-driven systems that suggest diagnostic, or treatment decisions are complex. In the event of an error, it is unclear who is liable: The physician, the healthcare institution, or the technology developer? It is imperative to institute transparent legal frameworks that delineate accountability to minimize any potential ambiguity of AI clinical use[36,37]. A more systematic examination of ethical challenges would underscore the critical importance of responsible AI governance and ethical safeguards, fostering trust among patients and healthcare providers.
ROBOTIC SURGERY
Cancer is recognized as a complex and systemic condition. Nowadays, a growing importance is being given to minimally invasive techniques, complemented by post-operative systemic therapies. This changing paradigm, compared to more extensive radical surgeries advocated in the 20th century, has enhanced patient outcomes[15]. Minimally invasive surgeries are a noteworthy advancement in GI surgical oncology, offering benefits not only in short-term and oncologic outcomes, but also in functional outcomes and quality of life[38,39]. Various early solid tumours have the potential for surgical cure, but surgical management of metastatic disease is challenging. Notwithstanding the clinical evidence supporting the surgical approach in specific settings, some divergences arise between evidence and real-world clinical practice.
The advent of robotic-assisted surgery has led in a new era of minimally invasive procedures, offering many advantages over traditional open and laparoscopy approaches. Among these are the three-dimensional, magnification and stable vision, empowering surgeons to execute complex procedures with minimal damage and better precision[40-43]. In the context of GI oncology, robotic surgery has become particularly valuable in oesophageal and pancreatic cancers, where precision is paramount[40,44,45]. Overall, these advances have been shown to translate to a reduced scarring, improved pain control, minimized blood loss, quicker recovery times, fewer wound-related complications and shorter length of hospital stay. Moreover, the robotic approach shows a decrease in conversion rates and fewer positive margins compared to the laparoscopic approach, when performed at experienced centers[40,46,47]. However, there is a significant gap in the literature about the lack of evidence supporting the benefits of robotic surgery in terms of patients related quality of life[14,40].
Besides safety concerns during robotic learning curve, the leading obstacle hampering broader implementation of this approach is the associated high cost[48,49]. Nevertheless, healthcare value is defined as quality of outcomes per currency unit applied to obtain those outcomes. The goal is on accomplishing the best patient’s outcomes at an acceptable cost. However, innovation itself embraces intrinsic value, mainly in technological developments that come before further revolutions. Leaping forward, the perfect match includes the combination of AI technology in future generations of robotic platforms, leading for automated surgery. Such kind of arrangements have the capacity to standardize techniques, improve quality, certify proficiency, and educate and foster next generation of surgeons.
PRECISION SURGERY
Since the early 1990s, it was recognized from a visionary perspective that it was worth integrating the available knowledge and concepts of molecular biology into the treatment of cancer. They were undoubtedly the first steps towards personalized clinical decision-making, leading to a less mutilating and tailored surgical approach, in opposition to the classic halstedian concept. Molecules are becoming an essential part of surgical oncology so that more and more targeted therapies can cure different cancers, even without surgery[50]. Additionally, pre-emptive surgery is now usually accepted to prevent cancer, as in patients with hereditary diffuse gastric cancer or with hereditary non-polyposis colon cancer and familial adenomatous polyposis, among others.
In surgical oncology, the concept of precision surgery has evolved considerably over the past century, comprising a deeper knowledge of the pathophysiology and disease features not immediately detectable to the human eye[36,51]. Precision surgery, or surgical precision medicine, refers to a surgical approach that applies advanced technologies, personalized data, and detailed understanding of individual patient features to optimize surgical outcomes. It aims to customize medical treatment to the specific genetic, environmental, and lifestyle factors for individual patients by tailoring therapeutic decisions. In this regard, the main purpose of precision surgery is to operate those patients most likely to really benefit from a specific surgical procedure.
NGS has enabled the rapid sequencing of tumour genomes, uncovering specific mutations that can be targeted with precision therapies. Personalized medicine, driven by these insights, allows for management tailored to the individual’s genetic makeup, then leading to more effective and less toxic treatments. There are many applications for NGS, including the management of infectious, dermatologic and oncologic diseases, and in genomic medicine[52]. Many GI cancers (e.g., colorectal, gastric, and pancreatic cancers) are treated based on specific genetic mutations (e.g., KRAS, BRAF). Genetic testing has become common in daily practice for GI cancers, but contrasting with lung or breast cancer, where the incidence of targeted gene abnormalities is low and non-standardized[34]. However, the impact on tumour response rate, patient outcomes, and survival of matched therapy has been already established. Lynch syndrome is one of the most familiar hereditary cancer syndromes that is caused by germline pathogenic variants in DNA mismatch repair. Nowadays, patients with KRAS/NRAS wild type are candidates for therapies targeting the epidermal growth factor receptor signalling. In contrast, patients with BRAFV600E mutation set up a RAS-independent activation of the mitogen-activated protein kinases pathway, which is a relevant prognostic biomarker in this setting[53]. Nevertheless, BRAF non-V600E mutated tumours tend to be resistant to BRAF inhibitors, with some response to epidermal growth factor receptor inhibition[54]. Currently, various mutations have been identified, and several precisely targeted drugs have been developed to treat metastatic colorectal cancer[53,55-57]. Some trials have shown a significant clinical benefit of the molecular-targeting combination therapy, in those patients[56,57]. Another emerging clinical application of NGS is the liquid biopsy. Traditional biopsies are invasive and often difficult to repeat, especially in patients with GI cancers where tumours may be in inaccessible areas like the pancreas or liver. Liquid biopsies enable the collection of tumour-related genetic material as circulating tumour DNA (ctDNA) from a simple blood sample, making it a minimally invasive method that can be performed multiple times over the course of treatment[58]. Liquid biopsies hold promises for early detection of GI cancers and for cancers that lack effective screening methods, such as pancreatic or colorectal cancers[59,60]. After treatment, liquid biopsies can be used to monitor minimal residual disease, detect early recurrence, and allows real-time monitoring of the tumour genetic profile, providing insights into new mutations or resistance mechanisms[61]. Furthermore, the application of ctDNA-based analysis may provide useful prognostic information. Higher levels of ctDNA are often associated with more aggressive disease or poorer prognosis, helping in risk stratification and supporting clinical decision-making of treatment intensity[58,62].
The challenges of NGS technology in a clinical setting is vast, including the complexity of data analysis, technological and operational limitations (e.g. Bioinformatics infrastructure), high cost and limited accessibility, tumour-specific features (e.g. multiplicity and complexity of mutations) and the lack of standardization and quality control[63,64]. Addressing these questions is essential for maximizing its clinical value. Solutions may include improving bioinformatics pipelines, reducing costs, standardizing protocols, and providing appropriate training for healthcare professionals.
Finally, NGS is a powerful tool used in microbiome analysis, particularly to understand the complexity of microbiota composition in health and disease. It has been even more useful in clinical settings to better figure out some postoperative complications like anastomotic leakage (AL). So far, some studies are aiming to define perioperative microbiota changes, and how these microbial alterations may be predictive of appropriate anastomotic healing. Certain microbial profiles are implicated in impairing wound healing or promoting inflammation and are associated with higher risks of AL and infection[65]. Research is ongoing to identify accurate microbial markers that can serve as early predictors of AL, allowing for personalized interventions to reduce the risk, such as probiotics, antibiotics, or other microbial-targeted therapies (e.g. faecal microbiota transplantation)[66]. Postoperative microbiome shifts, detectable by NGS, may also be useful in AL early detection, enabling a timely management.
THE CURRENT STATE-OF-THE-ART
Despite the confidence that these technological innovations will optimize oncological outcomes in patients with GI cancer, the current state-of-the-art in oncological treatment remains deeply entrenched in established protocols and clinical wisdom. Standard management including diverse chemotherapy and radiation therapy protocols, although associated with significant side effects, are still the most widely used treatment options for many GI cancers. Their efficacy, particularly in combination with surgery, has been demonstrated in large-scale clinical trials, providing a level of confidence that emerging therapies are still striving to accomplish. Multidisciplinary care models, also known as tumour board, have been developed since the second half of the last century. The current standard of care in GI oncology emphasizes a multidisciplinary approach, where different specialists work together closely sharing clinical decisions[35]. This holistic approach ensures that patients receive comprehensive care, and are thought to optimize patient outcomes, improving care performance. The institutional implementation of a tumour board has a significant impact on diagnosis, treatment strategies and processes. Multidisciplinary setting improved diagnosis and staging accuracy, changing the diagnostic reports in up to 35% of patients discussed[67]. Multidisciplinary discussion ensures a more appropriate management through preoperative reassessment of imaging and pathology results, warranting the most up-to-date treatment, and set up a structured oncological follow-up. With this approach, the likelihood of patients being offered non-palliative chemotherapy and radical surgery or radiotherapy increases[67]. In terms of oncological outcomes, it has reported a limited positive impact on local recurrence and distant metastasis rates after curative resection with two studies describing only a minimal positive impact on local recurrence rates for rectal cancer and incidence of metastases and remaining pelvic tumour after resection. Improvements in survival rates were also reported for colorectal cancer[67,68]. As far as waiting time is concerned, tumour board organization has been effective on the reduction of time from diagnosis to treatment, and on the accomplishment of an appropriate referral patterns[68]. One of the most important issues under discussion is exactly the advanced stage diagnoses and treatment backlog, considering the significant increase in GI cancer incidence. According to the scarce evidence available, only two-thirds of patients receive treatment within two months from an urgent referral and up to 96% national standard for receiving treatment within 4 weeks[14]. Another area of improvement was the development of a teaching environment for healthcare professionals. Greater integration of these patients’ research is also an increasing necessity to avoid a growing gap between outcomes related and those in actual clinical practice. Finally, these meetings increase the enrolment to tumour registries and stronger commitments to research and trials[14,67].
Subsequently, the focus will move to patient outcomes, particularly about the implementation of evidence-based practices and new technologies that have been demonstrated to result in tangible improvements. Some studies have supported that AI algorithms can significantly enhance polyp detection rates, potentially decreasing interval cancer cases by identifying high-risk lesions that might otherwise be missed[2]. A few comparative studies show the benefits of robotic-assisted surgeries on recovery times, surgical precision and postoperative outcomes[48]. These improvements in quality of life can be highlighted with statistics, such as shorter hospital stays and reduced readmission rates, underscoring the holistic benefits. Patient outcome metrics were evaluated, as for comparing overall survival and disease-free survival in patients treated with traditional methods vs those including robotics or AI-assisted technologies. Some recent studies have shown promising results, such as a 5%-10% increase in disease-free survival with the aid of robotic surgery in specific GI cancers like colorectal cancer[62]. Zureikat et al[69] have shown how robotic assistance, improving precision in pancreatic cancer resections, can reduce positive margin rates, potentially leading to fewer recurrences[8,59]. Advanced imaging techniques like narrow-band imaging and endoscopic ultrasound have shown promising results in improving early detection and staging accuracy for GI cancers, contributing to better survival rates[6]. Nonetheless, while robotic systems offer some technical advantages in GI cancer management, the evidence on long-term survival and recurrence rates remains uncertain[62].
THE DICHOTOMY: INTEGRATING INNOVATION WITH TRADITION
The need for a balanced approach
Moving forward, it is essential to adopt a balanced approach that integrates technological innovations with established clinical practices. There is a pressing need for robust clinical trials and real-world studies to validate the efficacy of these technologies across diverse settings. Furthermore, the development of cost-effective and accessible solutions is crucial to ensure that all patients, regardless of their socioeconomic status, can benefit from these advancements. Healthcare systems must invest in training and education to provide clinicians with proper skills to leverage these technologies effectively. The integration of multidisciplinary boards will be crucial in maximizing the potential of technological innovations. Whilst the privileges of technological advancements are obvious, their incorporation into clinical practice presents several threats. These include the expensive character of innovation and the budgeting for technology adoption, the need for extensive training, and the potential for inequity. The current state-of-the-art, based in robust evidence, offers a stable and reliable framework for GI cancer management. However, as technological advancements continue to evolve, there is a growing priority to bridge the gap between innovation and clinical evidence-based care. This requires a consistent strategic planning, by building sustainable workforce capacity, designing services to promote equity and improve outcomes, fixing the reimbursement system for cancer care and balancing the cancer research agenda.
CONCLUSION
In the future, to improve GI oncological outcomes, a delicate balance between technological advances adoption and evidence-based care should be chased. As we move forward, the question will be to harness the power of innovation while keeping up the highest standards of patient care. By fostering close collaboration between technology developers, clinicians and researchers, the translation from technology value into real-world improvements in patient outcomes is possible. The path ahead is complex, but with thoughtful integration and monitoring, the potential for transformative impact in GI oncological care is within reach.
Provenance and peer review: Invited article; Externally peer reviewed.
Peer-review model: Single blind
Specialty type: Oncology
Country of origin: Portugal
Peer-review report’s classification
Scientific Quality: Grade A, Grade B
Novelty: Grade B, Grade B
Creativity or Innovation: Grade A, Grade B
Scientific Significance: Grade B, Grade B
P-Reviewer: Guo ZY S-Editor: Fan M L-Editor: A P-Editor: Zhao S
