Innovative approaches in predicting outcomes for rectal neuroendocrine tumors

Abstract

Rectal neuroendocrine neoplasms pose significant challenges due to their varied presentations and prognoses. Traditional prognostic models, while useful, often fall short of accurately predicting clinical outcomes for these patients. This article discusses the development and implications of a novel prognostic tool, the GATIS score, which aims to enhance predictive accuracy and guide treatment strategies more effectively than current methods. Utilizing data from a large cohort and employing sophisticated statistical models, the GATIS score integrates clinical and pathological markers to provide a nuanced assessment of prognosis. We evaluate the potential of this score to transform clinical decision-making processes, its integration into current medical practices, and future directions for its development. The integration of genetic markers and other biomarkers could further refine its predictive power, highlighting the ongoing need for innovation in the management of rectal neuroendocrine neoplasms.

Key Words: Rectal neuroendocrine tumors; GATIS score; Tumor staging; Rectal neuroendocrine neoplasms; Survival prediction; Prognostic assessment; Biomarkers; Neuroendocrine carcinoma

Core Tip: The GATIS score is a new prognostic tool for rectal neuroendocrine neoplasms, enhancing accuracy by combining clinical and pathological factors. Its use in clinical decision-making could transform treatment approaches by tailoring them to patient risk levels, guiding management, and informing surveillance to detect early recurrence. This tool also aids in selecting clinical trial candidates, which could promote research and the development of new therapies.

TO THE EDITOR

Rectal neuroendocrine neoplasms (R-NENs) represent a unique and relatively rare category within the broader spectrum of NENs, originating from the neuroendocrine cells typically found in various organs like the gastrointestinal tract, pancreas, and lungs. These tumors are distinguished by their unique histopathological features and diverse biological behaviors, which significantly complicate their diagnosis and treatment. The range of R-NENs spans from well-differentiated neuroendocrine tumors (NETs), which generally have better prognoses, to the more aggressive poorly differentiated neuroendocrine carcinomas (NECs), leading to variable clinical outcomes and necessitating personalized treatment approaches. Despite their rarity, R-NENs can be categorized and graded using the World Health Organization (WHO) system for gastroenteropancreatic NENs, which assists in both prognostication and the formulation of treatment strategies[1,2].

Concurrently, R-NENs have been increasingly identified, largely due to advancements in screening and diagnostic methodologies. The estimated incidence of R-NENs is approximately 1.04 per 100000 individuals annually, although this is likely an underestimation since not all cases are systematically registered[3]. The prevalence of these tumors has risen over time, reflective of both an increasing incidence and their generally indolent nature, with a significant growth in prevalence observed among all NENs, demonstrating the impact of improved diagnostic capabilities[3,4].

The incidence of R-NENs has risen due to increased numbers of screening colonoscopies, leading to more incidental discoveries, enhanced endoscopic technologies for better detection and characterization, and heightened awareness among clinicians[2,5,6]. Data from the Surveillance, Epidemiology, and End Results database indicate a tenfold increase in the incidence rate in the United States from the 1970s to the 2000s, with current rates at about 1.1 per 100000 people[7]. In Germany, the incidence of R-NENs increased markedly between 1976 and 2006, with rates for males rising from 0.01 per 100000 in 1976-1978 to 0.26 per 100000 in 2004-2006, and for females from 0.01 to 0.24 per 100000 during the same period[8]. In Australia, the incidence of NETs was recorded at 0.244 per 100000 in 1996 and surged to 3.162 per 100000 in 2015, with the rectum being the most common site (29.65%)[9]. Similarly, in Taiwan, the incidence mirrored Australian figures, with the most prevalent sites being the rectum (29.65%)[10]. In England, NETs diagnosed through the English bowel cancer screening program showed rates of 29 rectal and 18 colonic NENs per 100000 colonoscopies[11].

Risk factors contributing to the development of R-NENs include ethnicity, with Japanese individuals at higher risk for larger tumors, and metabolic factors, such as high cholesterol and ferritin levels. Lifestyle choices like heavy alcohol consumption and demographic factors, such as younger age and male gender, also play roles in increasing susceptibility to R-NENs, alongside tumor characteristics, such as size and grade, that influence the risk of metastasis[12-15].

Treatment strategies for R-NENs vary based on tumor characteristics such as tumor size, grade, and stage, requiring tailored approaches for effective management. Small, well-differentiated R-NENs under 10 mm are typically addressed with endoscopic resection techniques, such as conventional polypectomy, endoscopic mucosal resection, and endoscopic submucosal dissection, which minimize the risk of invasion[2]. For larger tumors exceeding 20 mm, more invasive surgical options, such as trans anal excision, low anterior resection, or abdominoperineal resection are necessary due to the higher risk of spreading and deep tissue involvement[2]. Advanced or metastatic cases, especially poorly differentiated NECs, often require chemotherapy with platinum-based compounds, such as cisplatin or carboplatin combined with etoposide[16]. Additionally, targeted therapies such as everolimus and sunitinib are indicated in inoperable, advanced well-differentiated NETs by inhibiting specific growth pathways[17], while peptide receptor radionuclide therapy targets tumors expressing somatostatin receptors to control growth and manage symptoms effectively[18]. Regular post-treatment surveillance through endoscopic and imaging exams is crucial, especially for intermediate-sized tumors, to monitor for recurrence. The comprehensive management of R-NENs typically involves a multidisciplinary team, with the goal of delivering personalized and thorough patient care[2].

The prognostic assessment of R-NENs utilizes a variety of systems and methods to effectively predict clinical outcomes based on tumor characteristics. The WHO classification system categorizes R-NENs into G1, G2, or G3 based on tumor grade and the Ki-67 proliferation index, where higher grades typically indicate a poorer prognosis[19]. Tumor size is another critical prognostic factor; smaller tumors (< 1 cm) often have a favorable prognosis, whereas those larger than 2 cm or with invasive features usually predict worse outcomes[20]. Advanced staging systems, such as those from the American Joint Committee on Cancer and the European Neuroendocrine Tumor Society, further refine prognosis by classifying tumors based on size, lymph node involvement, and the presence of metastases[21]. Additionally, imaging techniques, such as positron emission tomography imaging with 18F-fluorodeoxyglucose, help in stratifying risk by detecting the metabolic activity of tumors, with higher uptake indicating poorer prognoses[22]. Moreover, nomograms and tools, such as the Nodal involvement and Other Variables for Assessment of Risk in rectal neuroendocrine tumors score, provide more individualized predictions by incorporating factors such as age, tumor size, grade, and vascular infiltration[23,24].

On the other hand, predicting the prognosis of R-NENs presents multiple challenges due to the rarity and heterogeneity of these tumors, which complicates gathering large patient cohorts essential for robust studies and leads to limited statistical power and generalizability[5]. These tumors display a range of biological behaviors from well-differentiated NETs to aggressive NECs, each with different prognostic implications[25]. The lack of a standardized prognostic model specifically for R-NENs frequently means that models developed for other NENs are inappropriately applied, which may not fully capture the unique aspects of R-NENs. Additionally, while biomarkers, such as the Ki-67 index and molecular profiling tools, show potential, their reliability and applicability in R-NENs are still under scrutiny due to inconsistent performance across studies[26]. Challenges also arise from traditional imaging and staging methods that may not accurately reflect disease progression, highlighting the need for the integration of advanced imaging techniques and updated staging systems that require further validation[21].

GATIS SCORE

The GATIS score combines various clinical and pathological factors to predict the prognosis of R-NENs. These five key factors include tumor grade, tumor stage, tumor size, age, and the prognostic nutritional index, which is calculated as serum albumin plus five times the peripheral blood lymphocyte count. Together, these factors are used to assess overall and progression-free survival in patients with R-NENs. The WHO 2010 classification system was used to categorize R-NENs, while in the American Joint Committee on Cancer 8^th edition, tumor, nodes, and metastases (TNM) staging criteria were applied for NET, NEC, and mixed adenoneuroendocrine carcinoma staging[27].

The GATIS score could facilitate tailored clinical decision-making in managing R-NET. It enables personalized treatment intensities and surveillance protocols by categorizing patients based on their prognostic risks. Additionally, the score streamlines the selection of clinical trial participants, fostering homogeneity and enhancing the reliability of trial outcomes. This targeted approach accelerates the development of new therapies by identifying patient groups most likely to benefit from novel treatments.

The development of the GATIS score could be considered a pivotal advancement in the prognostic assessment of R-NENs, addressing significant gaps in current prognostic tools, such as the WHO classification and TNM staging systems, which often fall short in predicting clinical outcomes for these tumors. Various classification systems have been proposed for R-NENs. The TNM system lacks adequate predictive accuracy for lymph node metastasis, a crucial prognostic factor for RNETs[28]. While the TNM staging system primarily focuses on anatomical staging, it fails to incorporate the critical prognostic implications of tumor grading, such as G1, G2, and G3 classifications. Multivariate analyses indicate that tumor grade and size are often more significant independent predictors of prognosis than TNM staging alone[29].

The WHO classification system groups tumors into grades (G1, G2, G3). However, the evaluation criteria remain controversial[30]. These criteria often fail to adequately differentiate between tumors of similar grades that exhibit distinctly different clinical behaviors. For instance, L-cell-type tumors are usually considered less aggressive, yet they can exhibit malignant behaviors, a variability not fully captured by the current WHO classification[31]. Furthermore, the system does not consistently integrate newer biomarkers, such as chromogranin A or cyclin A, which could improve predictions of metastatic potential beyond what is possible with the Ki-67 index alone[32]. Additionally, while the WHO grades tumors based on mitotic count and the Ki-67 index, it overlooks the significant risks associated with small tumors (i.e., under 1 cm), that still possess metastatic capabilities[33].

One suggestion is a revised lymph node staging system that considers the number of positive lymph nodes (N0/N1a/N1b/N2), rather than the traditional binary N0/N1 classification, for more accurate prognostic predictions[34]. Additionally, histological grade and tumor size have been identified as important prognostic factors, with a non-L-cell immunophenotype and tumor size greater than 1 cm linked to more aggressive behavior and worse prognosis, highlighting the need for staging systems to include both tumor size and L-cell phenotype for better predictive accuracy[31].

The GATIS score predictive model has been derived from a comprehensive dataset of 1408 patients from 17 major medical centers, enhancing its statistical robustness and generalizability across various clinical settings[27]. The GATIS score leverages advanced statistical methods, including random forest and Cox proportional hazard models, to manage complex data interactions and accurately predict patient prognosis. This methodological sophistication allows for a refined evaluation of diverse patient demographics and tumor characteristics, which is crucial for tailoring treatment plans and improving patient management. The score’s ability to distinguish between patients needing aggressive vs conservative treatment approaches directly influences outcomes, potentially enhancing both survival rates and quality of life.

Furthermore, the introduction of the GATIS score represents a significant step towards the application of personalized medicine to oncology, particularly for managing R-NENs. By integrating modern biomarkers and specific tumor characteristics, the GATIS score provides a more nuanced and predictive approach than traditional models. Its validation through internal and external cohorts confirms its accuracy and reliability, setting a foundation for its clinical application and further research. The ongoing development and refinement of the GATIS score, including the potential incorporation of additional biomarkers and clinical factors, promise to enhance its utility in diverse clinical scenarios. This model not only facilitates better patient stratification in clinical trials but also aids in the precise evaluation of treatment efficacy, accelerating the development of new therapies and advancing the understanding of R-NENs. The GATIS score thus stands as a benchmark for future prognostic models for the diagnosis of R-NENs, emphasizing the importance of tailored approaches and continued innovation in the management of NETs.

Comparative studies between China and the United States reveal notable differences in the incidence and outcomes of R-NENs. The age-standardized rate of incidence in China is reported at 0.16 per 100000, while it is higher in the United States at 0.99 per 100000. Despite a lower incidence, survival rates in China are also lower, with a five-year relative survival rate of 69.8% (confidence interval: 61.5-78.1), compared to 89.6% (confidence interval: 87.5-91.7) in the United States[35]. These disparities underline significant differences in the outcomes and possibly in the treatment approaches for R-NENs between the two countries. Given these variations, it is recommended to validate the GATIS score in the United States to ensure its global applicability and relevance.

LIMITATIONS

The study on the GATIS score for predicting the prognosis of R-NENs showcases notable methodological strengths but also carries limitations and potential biases that might affect its generalizability and accuracy. As a retrospective analysis, it is constrained by the inherent limitations of historical medical records, including possible selection and information biases due to missing or inaccurately recorded data. Since the validation was conducted entirely within China, its findings may not be easily generalizable to other populations due to differences in genetic, dietary, and environmental factors that can influence disease progression and patient outcomes. Additionally, the sophisticated statistical methods employed, such as random forest and Cox proportional hazards models, while robust, risk overfitting when handling numerous predictors, which could diminish the model’s accuracy outside of the original dataset.

Although the GATIS score underwent both internal and external validation, the absence of prospective validation in independent cohorts means that further validation is needed to fully establish its utility in clinical settings. The exclusion of patients with simultaneous or metachronous malignancies could introduce selection bias, limiting the score’s applicability across all R-NEN patients. Moreover, the study’s follow-up duration may not have been sufficient to capture long-term outcomes for patients with slower disease progression, potentially affecting the accuracy of the predicted outcomes. Lastly, the GATIS score’s heavy reliance on quantifiable clinical and pathological markers, without incorporating molecular and genetic markers, might restrict the depth of the prognosis assessment, underscoring the need for a more comprehensive approach in future studies.

PROPOSED FUTURE RESEARCH DIRECTIONS

To enhance the predictive accuracy and clinical utility of the GATIS score for prognostication of R-NENs, future research should focus on integrating genetic and molecular biomarkers, employing advanced imaging techniques, and ensuring comprehensive validation. Incorporating genetic profiling, including mutations in key genes, TP53, APC, KRAS, BRAF, RB1, CDKN2A, and PTEN[36], along with molecular biomarkers, such as hormone levels and protein expression profiles, can provide deeper insights into tumor pathology and influence prognostic accuracy[37]. Advanced imaging, such as radiomics, could further refine prognostic assessments by revealing detailed phenotypic information and sub-visual features of the tumor environment. The application of the GATIS score to multi-national cohorts and its prospective validation through longitudinal studies would adapt the model to diverse genetic backgrounds and clinical practices, enhancing its global applicability and its ability to predict treatment responses. Future studies using the GATIS score could aid in stratifying patients, allowing researchers to more effectively identify subgroups with similar prognostic characteristics who may respond differently to various treatment regimens. Collaborative multidisciplinary efforts across oncology, genetics, bioinformatics, and radiology are crucial for refining the score, and addressing the economic and ethical implications of its implementation. These considerations will be key for its integration into clinical practice.

GATIS SCORE COULD BE INTEGRATED INTO CLINICAL DECISION-MAKING PROCESSES

It appears that the integration of the GATIS score into clinical decision-making processes for managing R-NENs could potentially be transformative, although this remains speculative. At diagnosis, the GATIS score attempts to stratify patients by risk, which might influence the intensity and type of treatment necessary. This could promote more aggressive approaches for high-risk patients while suggesting conservative management for those with lower scores. Such stratification may aid in selecting appropriate surgical, chemotherapeutic, radiotherapeutic, and novel treatments, such as peptide receptor radionuclide therapy, as well as in timing these interventions to optimize patient outcomes. Additionally, by dictating surveillance strategies, the GATIS score might adjust the frequency and intensity of follow-ups to attempt early detection of disease recurrence, thereby facilitating timely interventions and potentially improving overall treatment efficacy. Furthermore, it could enhance team coordination within multidisciplinary oncology groups by providing a common framework that improves collaborative decision-making and serves as an educational tool for aligning treatment plans and understanding patient prognosis. Identifying factors with significant impact on prognosis could better facilitate the decision-making process and enhance patient communication, allowing for more informed discussions about treatment options and expectations. For instance, a young patient with a low-risk, small tumor (1 cm, T1, G1) and a low GATIS score might be recommended for surgical excision. Conversely, an older patient with poor nutritional status and a high-risk, large tumor (> 2 cm, T3, G3) might benefit from neoadjuvant chemotherapy to address aggressive disease characteristics and potential metastases, as indicated by a high GATIS score. It could also assist in the selection of candidates for clinical trials, thereby advancing research and the development of new therapies for R-NENs.

CONCLUSION

The GATIS score represents a significant advancement in the prognostic assessment of R-NENs, offering enhanced predictive accuracy by incorporating key clinical and pathological factors, such as tumor grade, tumor stage, tumor size, age, and prognostic nutritional index. It addresses the limitations of traditional staging systems, such as WHO and TNM, by providing a more comprehensive assessment framework. While initially validated in Chinese cohorts, further international studies are essential to confirm its global applicability. This tool holds promise for refining treatment strategies and emphasizing the necessity for personalized medical interventions in oncology.