Predicting early recurrence in resectable rectal cancer

Abstract

We provide an editorial of recent findings on early recurrence (ER) in rectal cancer (RC), focusing on the study on ER of resectable RC by Tsai et al. The study established an 8-month recurrence-free survival cut-off for differentiating ER from late recurrence, with implications for postrecurrence survival and overall survival. This offers not only a valuable timeframe for enhancing surveillance strategies in patients at higher risk, especially those who have undergone neoadjuvant chemoradiotherapy (CRT), but also raises questions about its applicability across different populations. Furthermore, the article suggests that while CRT is very effective in reducing locoregional recurrence, this treatment alone may not fully address the overall risk of ER. The authors advocate for personalized risk assessment and intensive surveillance during the postoperative period to improve outcomes, particularly in the first year. Future research should focus on larger, multicenter studies and incorporate advanced diagnostic techniques with genetic and molecular biomarkers to enhance prediction and management of ER. The ultimate goal is to refine treatment and surveillance strategies to improve survival and quality of life for patients with RC.

Key Words: Resectable rectal cancer; Early recurrence; Chemoradiotherapy; Recurrence-free survival; Advanced imaging; Molecular biomarkers

Core Tip: The identification of an 8-month cut-off to distinguish early recurrence (ER) from late recurrence in rectal cancer emphasizes the importance of targeted surveillance in this period. Factors such as advanced age, preoperative chemoradiotherapy, and longer hospital stays highlight patient subsets that require intensive follow-up. Enhanced surveillance protocols with advanced imaging and molecular biomarkers could better manage recurrence and define risk groups for ER, improving treatment strategies and thus survival outcomes and patients’ quality of life.

INTRODUCTION

Rectal cancer (RC) accounts for about 30% of colorectal cancer cases and is one of the leading causes of oncological morbidity and mortality worldwide[1]. The therapeutic approach to resectable RC has evolved significantly over the past decades with the introduction of neoadjuvant therapies aimed at improving oncological outcomes and minimizing the adverse effects associated with treatment[2]. Among these strategies, neoadjuvant radiotherapy (RT) and concurrent or sequential neoadjuvant chemoradiotherapy (CRT) have stood out, providing proven benefits in terms of tumor control, reduction in recurrence rates, and higher rates of complete total mesorectal excision (TME). These outcomes were further improved with the introduction of total neoadjuvant therapy (TNT), reaching pathological complete response (pCR) rates up to 37%[3].

The recent study by Tsai et al[4] published in the World Journal of Gastrointestinal Surgery offers valuable insights into the timing and risk factors for early recurrence (ER) in patients with resectable RC. The study’s retrospective analysis, encompassing 131 patients over nearly a decade, identifies a recurrence-free survival (RFS) cut-off of 8 months, differentiating ER from late recurrence (LR). This analysis also highlights critical prognostic factors such as age, preoperative CRT, colostomy creation, and length of hospital stay. While the study contributes significantly to understanding ER, several aspects require further exploration, particularly regarding the integration of these findings into clinical practice and the broader implications for patient management.

KEY FINDINGS AND IMPLICATIONS

The study underscores the importance of recognizing ER as a distinct clinical entity. The identification of an optimal RFS cut-off point at 8 months provides a critical timeframe for clinicians to focus surveillance efforts. We highlight the importance of predicting which cases are most likely to experience ER.

CHALLENGES IN DEFINING ER

The study’s single-center design and limited sample size further suggest that these findings should be validated in larger, multicenter cohorts to ensure their applicability in broader clinical contexts.

A major limitation of Tsai et al’s study is the lack of distinction between local recurrence and distant recurrence[4]. Therefore, calculating a single cut-off for ER from a sample that, besides being too small, is also too heterogeneous, with a poorly stratified analysis (including patients with both localized and locally advanced disease and with both local and distant recurrences) could, in our opinion, compromise all subsequent analyses and the generalizability of the findings. The values obtained for median postrecurrence survival (PRS) of 1.4 months (for ER) and 2.9 months (for LR) reinforces this heterogeneity and the RFS cut-off obtained for ER of 8 months, which is lower than the estimates provided by current evidence, could be a sign of this compromise.

The study employed a receiver operating characteristic curve to define ER, setting it apart from previous studies that arbitrarily assigned cut-off points ranging from 1 year to 5 years post-surgery. While this interesting approach addresses a significant gap in the literature, it also raises questions about the generalizability of the 8-month cut-off across different populations and treatment settings.

Nevertheless, defining an 8-month period as the cut-off for ER provides a useful guide for clinical practice, allowing better risk stratification and implementation of more effective surveillance strategies. The lack of correlation between RFS and PRS and the existence of one between ER and PRS may reinforce the vital prognostic importance of ER prediction and the applicability of this specific cut-off value. Future studies should focus on optimizing treatment and follow-up strategies for these patients with the aim of improving overall survival and quality of life.

RISK FACTORS AND THEIR CLINICAL RELEVANCE

The identification of specific risk factors of ER, namely, age over 70 years, preoperative CRT, colostomy creation, and prolonged hospital stay, provides valuable insights for clinicians. However, these findings also highlight the complex interplay between these factors and ER.

First, some of these factors obviously merge and correlate, as is the example of advanced age and longer hospital stay. Moreover, the identification of colostomy creation as a risk factor for ER opens the discussion on the best therapeutic approach to adopt in patients with low rectal tumors. The fact that the colostomy does not influence the survival outcomes highlights the importance of other targeted or systemic therapeutic strategies adopted in ER cases.

In light of current evidence, we know that neoadjuvant RT, used prior to TME, has demonstrated substantial benefits, such as tumor size reduction, increased resectability rates, decreased local recurrence, and improved survival[5,6]. Neoadjuvant concurrent CRT, which combines RT with chemotherapy (ChT) agents such as 5-fluorouracil (5-FU) or capecitabine, has demonstrated superior efficacy compared to RT alone with better rates of pCR and increased survival, although with greater toxicity[7,8]. Studies, such as CAO/ARO/AIO-94, National Surgical Adjuvant Breast and Bowel Project R-03, and German Rectal Study have shown that preoperative CRT, compared to postoperative CRT, results in better local control rates and lower toxicity indices[9-13]. Preoperative CRT improves downstaging, the rates of pCR and disease-free survival (DFS), and decreases the incidence of positive surgical margins and LR with better sphincter preservation rates and less late toxicity[9-13]. More recently, novel regimens of ChT associated with RT (concurrently and/or sequentially) in the so called TNT have demonstrated higher rates of organ preservation for patients with locally advanced tumors, offering better sphincter preservation rates and a lower need for permanent colostomies[3,14-20]. TNT trials, such as RAPIDO, PRODIGE 23 and STELLAR, have evaluated the efficacy of intensified neoadjuvant CRT regimens, showing further improvements in overall survival, DFS and pCR[14-20].

In fact, the association between undergoing CRT and the occurrence of ER emphasizes the need for more rigorous surveillance in these patients. However, tumor biology and aggressiveness in locally advanced disease should prevent us from identifying CRT as an independent risk factor for ER. Furthermore, it would have been relevant to conduct a subgroup analysis for both localized and locally advanced tumors and for local and distant recurrence in patients who have completed prior CRT.

Besides, the global association between preoperative CRT and ER suggests that while this treatment modality is clearly effective in reducing local recurrence and improving survival outcomes, it may not sufficiently mitigate the overall risk of ER in all patients, most likely due to nonregional/distant recurrence.

These findings reinforce the need for a careful and personalized risk assessment in patients undergoing neoadjuvant treatments, particularly those with additional risk factors such as advanced age and the need for colostomy. Intensive surveillance after surgery, especially in the first year, may be crucial for the early detection of recurrences, allowing for timely interventions that could improve clinical outcomes.

PREDICTION OF TREATMENT RESPONSE

The crucial period when enhanced surveillance is required and the critical importance of predicting which cases are most likely to experience an ER, are the two fundamental ideas that we underscore.

As aforementioned, TNT has increased pCR rates compared to conventional treatment. The incorporation of molecular biomarkers, genetic profiling, and advanced imaging techniques could enhance tumor response and ER prediction, enabling more personalized therapeutic approaches and potentially preventing unnecessary surgical interventions. However, reliable biomarkers are lacking and current evidence suggests that there is no single biomarker capable of clearly predict disease progression[3].

Biomarkers are widely studied due to their accessibility and potential prognostic value. One of the most studied are hematological biomarkers. Higher values of neutrophil-to-lymphocyte ratio, platelet-to-lymphocyte ratio and monocyte-to-lymphocyte ratio (MLR) are generally associated with poor prognosis and lower pCR rates, but with limited standalone predictive utility. MLR, in particular, has been identified as one of the most robust prognostic indicators, especially when combined with magnetic resonance imaging (MRI) findings[3].

Carcinoembryonic antigen (CEA) is widely used for monitoring, but it has limited sensitivity, especially in early-stage disease. Circulating tumor DNA has higher sensitivity and specificity than CEA and thus can better identify minimal residual disease and recurrence after CRT and surgery. However, its ability to predict pCR remains unproven[3].

KRAS and BRAF mutations (particularly BRAF V600E) correlate with poor prognosis and high recurrence rates and predict response to anti-epidermal growth factor receptor therapies. A lower MLR and positive peripheral nerve infiltration are strongly associated with ER in BRAF-mutant colorectal cancer[3,21-23]. Microsatellite instability and mismatch repair deficiency predict resistance to 5-FU-based ChT (FOLFOX), but are associated with better responses to checkpoint inhibitor immunotherapy, potentially eliminating the need for CRT or surgery in selected patients[3,21,22]. Both excision repair cross-complementing 1 and 2 overexpression are associated with a poor response to FOLFOX-based concurrent CRT[22]. Finally, the MRE11/RAD50/NBS1 (MRN) complex is involved in DNA repair and higher expression of MRN proteins may predict radiosensitivity, offering a potential biomarker for RT response[3].

In fact, biomarker combinations, rather than single markers, provide the most reliable predictive value. Integrating functional MRI findings with biomolecular markers could enhance precision in treatment stratification.

MRI plays a crucial role in staging and assessing treatment response. Functional MRI sequences, such as diffusion-weighted imaging and apparent diffusion coefficient, have been investigated for their ability to predict response to neoadjuvant therapy, with interesting results. Perfusion MRI, which assesses tumor microcirculation and vascular permeability, has shown promise in predicting tumor invasiveness and response, particularly through late slope parameter. The magnetic resonance tumor regression grade criteria utilize T2-weighted MRI to accurately correlate imaging with pathological tumor regression after neoadjuvant treatment, classifying responses into five grades. MRI-detected extramural vascular invasion and circumferential resection margin are well-established markers of poor prognosis, but its correlation with hematological biomarkers remains under investigation[3,24].

Additionally, when incorporated into emerging approaches like radiomics and artificial intelligence-driven multi-omics models, these indicators could significantly improve response prediction accuracy and treatment adaptation in the future. Machine learning has demonstrated strong performance, with some models achieving area under the curve values up to 0.99. Deep learning neural networks have achieved 80% accuracy in predicting complete response after neoadjuvant CRT, surpassing others. Radiomics-based nomograms further enhance predictions[3,23-25].

FUTURE RESEARCH AND CLINICAL PRACTICE

These findings support the development of personalized, noninvasive treatment strategies for improved colorectal cancer management. Additionally, while intense postoperative surveillance is warranted, the cost-effectiveness and psychological impact of such an approach must be carefully weighed against its potential benefits.

CONCLUSION

Tsai et al’s study provides a critical foundation for understanding the dynamics of ER in resectable RC[4]. However, translating these findings into clinical practice requires an adapted approach that considers patient and tumor-specific factors, as well as the need for personalized surveillance protocols. As we refine our understanding of ER, the ultimate goal remains to enhance survival outcomes and quality of life for all patients with RC.