An individualized treatment approach is necessary to improve survival and quality of life in rectal cancer. Tools to stratify patients are currently missing. Circulating tumor DNA seems to be a good candidate, but results are sparse, conflicting and characterized by lack of a uniform approach. The ctDNA status may be of importance to the long-term prognosis, but the area of research is new and is short of dedicated studies. There is an obvious need for standardization in ctDNA research, and the issue should be addressed in future research.
1. Introduction
The treatment of locally advanced rectal cancer (LARC) has evolved from surgery alone to experimental trimodality treatment comprising neoadjuvant chemoradiotherapy (nCRT) followed by surgery. Modifications of radiation doses and treatment schedules have been presented with different rates of pathological response but without improvement of long-term survival
[1][2][3][4][5]. Complete pathological response remains an important prognostic factor but is a poor surrogate endpoint for overall survival
[6].
The achievement of complete pathological response after chemoradiation suggested the possibility of organ preservation. A selected group of rectal cancer patients achieving complete tumor regression, termed clinical complete response (cCR) have been treated according to this approach
[7]. Several studies have reported excellent oncological outcome and high quality of life
[8][9][10][11], but a relatively high fraction will recur during follow-up. It remains unclear how patients should be selected for surgery versus observation after chemoradiation. Total neoadjuvant therapy is a new approach in rectal cancer, but the concept needs further validation before it comes to clinical utility
[12][13].
The treatment of rectal cancer calls for new markers, which alone or in combination can guide a more individualized treatment approach.
Circulating tumor DNA (ctDNA) is a new biomarker that has gained rapidly increasing interest. In many aspects, it seems to fulfill the criteria of an ideal tumor marker. It is easily accessible for repeated measurement with little discomfort to the patient, reflects important biological characteristics and seems to overcome the issue of tumor heterogeneity
[14][15]. There is a steep increase in the number of publications, witnessing a considerable clinical interest. In colon cancer, several prospective trials are investigating if ctDNA can identify patients with non-metastatic colon cancer, with high risk of recurrence, that may benefit from adjuvant chemotherapy
[16][17][18][19]. However, in rectal cancer, the influence of ctDNA on treatment decisions is still negligible. The diverging results and lack of standardization pose a major problem and impede further development. The situation calls for a comprehensive evaluation of the current literature.
2. Current Insights
The current literature points toward ctDNA as a promising tumor marker in clinical oncology. In rectal cancer, ctDNA may result in an individualized treatment approach. Reviews published during the recent years have tried to outline the role of ctDNA in rectal cancer, but none of them have addressed the quality of the included studies
[20][21].
Bias was assessed with the QUIPS tool. Six studies were classified as having low bias, two with moderate risk and one with high risk of bias. The developers of the QUIPS suggest that studies with high risk of bias should be omitted from further analysis. Based on this statement, only six out of nine studies in the research would be eligible for the evaluation of ctDNA. Since the population in the research is small and the results are reported narratively, the researchers present all results from the analysis. However, this serves as a clear indicator of the necessity of a quality assessment for future reviews.
There is no evaluation of bias in papers reviewing ctDNA in rectal cancer. The lack of quality appraisal is a problem in many reviews. A literature search of Hayden et al. revealed a rising number of published reviews and a consistent problem with the quality assessment. Almost 50% of the articles generated from their search were excluded because of lacking quality assessment. For the remaining articles, the evaluation varied and was often inadequate. Because of this gap, they developed QUIPS to assess validity and bias in studies investigating prognostic factors
[22]. This is the only tool specifically developed for quality evaluation of prognostic studies. Although the tool indicates potential problems with only three out of nine studies, it should be used cautiously, as it does not address all problems, i.e., comparison of methodological issues.
Riley et al. suggest a guide that should be applied when conducting systematic reviews and meta-analyses of prognostic studies
[23]. In the future, this systematic approach will lead to fewer reviews with comprehensive analysis of the included literature. As a result, the role of different biomarkers may be clarified faster, and unnecessary research can be omitted.
The REMARK checklist was developed to ensure consistent reporting
[24][25]. In spite of the introduction of the checklist, results have been discouraging with no improvement in reporting
[26]. This is supported by the researchers' findings with only one study reporting the use of the checklist. Hence, there is a risk of continuous research in markers without real progress as to clinical utility.
Another issue is the lack of standardized methods for ctDNA detection and reporting, as this impairs comparison and generalization of results. This is indeed what the researchers see in this entry. Characteristics and features of the tumor, which are not fully understood, can aggravate this problem along with other problems as illustrated in the literature
[27][28][29][30][31][32][33]. The discussion of the methodology is outside the scope of this entry, but a comprehensive entry has addressed the current challenges and suggests possible solutions which should be considered in future research
[34].
The predictive value of ctDNA varied among the included studies. Detectable ctDNA at baseline does not seem to be a predictor of treatment response. Only two of seven studies report the presence of ctDNA after CRT and a poor response as evaluated by MR or the surgical specimen, respectively
[29][35]. These results indicate that ctDNA currently cannot be used to select treatment, i.e., organ preservation or surgery. In general, the results should be carefully considered, since none of the studies report power calculations. This is likely a result of the retrospective research design. For the prospective studies, the ctDNA analyses are exploratory, and it may be rather difficult to make proper assumptions regarding the power calculations. Post hoc power calculations are generally not recommended
[36]. Hence, the literature is short of dedicated prospective trials exploring the clinical aspects of ctDNA in rectal cancer. However, in colon cancer, several prospective trials are investigating if ctDNA can predict recurrence and if ctDNA-guided treatment can improve survival
[17][18][34][37]. Circulating tumor DNA seems promising in other types of cancers, but results are limited by the relatively small size of the trials
[38][39][40].
For the survival outcomes, all of the studies found a correlation between survival outcome and ctDNA at a certain time point. For the baseline measurements, three out of nine reported an association between detectable ctDNA and worse survival. For the pre- and post-surgery analyses, four out of six and five out of five, respectively, found an association between detectable ctDNA and worse outcome. The persistence of ctDNA throughout the course of treatment seems to be a poor prognostic marker. The reported increasing HR at each time point supports this hypothesis. Especially detectable postoperative ctDNA seems to correlate with relapse. Similar results are reported in the treatment of colon cancer and malignant melanoma
[38][41].
At this point, it is difficult to conclude as to the prognostic importance of ctDNA in rectal cancer as most studies are small, the statistical assumptions are dubious and the follow-up period is rather short in some of the included studies. There is a substantial risk of both missing a correlation and finding one that would not be present if a longer follow-up period were available. In a review assessing trial extensions using routinely collected data, the authors explored if there were any beneficial or harmful events in the period beyond the original trial time period. They showed that in nearly 50% of the trials, a benefit was observed with extended follow-up, and in nearly 30% of these, the benefits were only significant for this period. Loss of significant benefits was only seen in 7% of the studies
[42]. These results highlight the importance of a sufficiently long follow-up period in terms of evaluating interventions, new diagnostic or prognostic tools. For rectal cancer, most local regrowths appear within the first two years, while metastatic disease may appear several years later. Therefore, both the follow-up period and the evaluation of the endpoint should be selected carefully prior to trial initiation
[43]. Studies with large cohorts and long follow-up are time consuming and expensive, but they are necessary in terms of clarifying the real value of biomarkers. A surrogate endpoint would be ideal to select optimal treatment and reduce costs and follow-up in rectal cancer. The study published by Jakobsen et al. indicates that ctDNA response may serve as a surrogate marker for OS in several cancers
[44]. Organ preservation in rectal cancer as standard treatment could be a possibility if ctDNA response proved to be a reliable surrogate marker for oncological outcome.