According to Colorectal Cancer Statistics 2023, the majority of CRC patients are elderly, and the number is continuing to climb due to advanced aging (5, 18). In this study, over thirty-percent of the colorectal cancer patients were ≥ 75 years old, so colorectal cancer patients aged ≥ 75 years were defined as ECRC. Many independent risk factors were shown to be positively related to the early death in ECRC patients, including Caucasian race, high tumour grade, higher T stage, higher N stage, larger tumour size and distant metastases (bone, lung, liver and brain). Surgery, chemotherapy, as well as radiotherapy were negatively related to premature death. This study is the first to look into the high risk factors for early death in ECRC patients. Based on these results, this study further developed a risk prediction model to assess the risk of early death in ECRC patients. The model was presented as a nomogram.This study can further investigate the incidence of early death in ECRC patients and guide clinical treatment.
In some regions in Asia, such as China and Japan, ECRC patients have seen a significant rise over the past two decades. According to one study, China has the fastest-rising incidence rate of CRC patients aged 65 and over, with a projected 311/100,000 by 2030, three times the number in 1988(19). In recent years, there have been many changes in how clinicians treat CRC. In the past, the treatment options for CRC were mainly based on surgery and chemoradiotherapy. In recent years, thanks to developments in histopathology and genetics, the treatment of CRC has moved to the genetic level. Targeted therapy and immunotherapy have rapidly developed on this basis, significantly extending patients' survival. However, due to the high incidence and malignancy of colorectal cancer, many patients still die from its progression (20). In the present cohort of 16,512 ECRC patients, 20.9% of those patients died early. Of these early deaths, 69.3% died of cancer and 30.7% were not cancer-specific deaths. Accumulating research has demonstrated the factors affecting the prognosis of metastatic colorectal cancer, while few focused on how age affects CRC patients. The treatment of ECRC patients is a modern challenge for personalized treatment, and developing a treatment plan for ECRC patients needs to take more factors into consideration(21, 22).
According to the results of this study, surgery, chemotherapy as well as radiotherapy were all key prognostic factors for early death in ECRC patients. Nowadays, surgery is the primary curative option for CRC(23). In previous studies, the prognosis of surgery differed between elderly and young patients because of more underlying disease, slower postoperative healing, minor and severe postoperative complications, and inability to tolerate surgery. Age ≥ 65 years is considered as an independent risk factor for death in gastrointestinal surgery (24). Interestingly, Samuelsson discovered that older cancer patients seem to be more suitable for elective surgery and more tolerant of surgical stress than generally assumed (25). Turri et al. also supported that colorectal surgery could be given even to older patients if postoperative mortality was acceptable (26). Consistently, the nomograms for predicting early mortality suggested that surgery could provide a significant survival advantage for ECRC patients in this study, indicating that the patients who received surgical intervention were at lower risk of death. In addition, high-risk elderly and frail patients should undergo further preoperative assessment.
The use or absence of chemotherapy was a major risk factor for early death in patients with ECRC, suggesting that chemotherapy may prolong survival and improve prognosis in ECRC patients. In recent years, the treatment of colorectal cancer has entered a new field thanks to the proposal of various chemotherapy regimens and clinical trials (23). Previous studies have found that ECRC patients can derive equal benefit from chemotherapy as young patients, with no significant increased toxicity (27). It has also been shown that most high-risk patients did not benefit from adjuvant chemotherapy (28).Thus, in frail elderly patients, especially those with declining organ function and preexisting comorbidities, careful chemotherapy regimens must be developed and adverse events carefully monitored(29). Nomograms developed in this study will be capable of providing clinicians with a reference for their decision on whether chemotherapy is needed for elderly patients.
Radiotherapy, chemotherapy, and surgical resection are often used in combination to confront CRC and associated metastases(30). Patients who receive radiotherapy have a lower risk of local recurrence (31). This study also found that ECRC patients who received radiotherapy were less likely to experience early death. However, one study showed that older colorectal cancer patients received radiation therapy nearly 70% less frequently than younger patients, although evidence suggested that they may benefit as well(7). Some elderly patients might not be suitable for radiotherapy because of poorer underlying physical conditions and more comorbidities that make them unable to tolerate the side effects of radiotherapy. However, this does not explain why elderly patients receive radiotherapy at a much lower rate than younger patients. For now, the potential of radiotherapy for the therapeutic effect on ECRC patients has not been fully exploited. This study encourages the provision of radiotherapy to elderly patients who can tolerate it, thus it is necessary for clinicians to consider radiotherapy more in the treatment of ECRC patients in the future.
Previous studies have recognised distant metastasis as a main driver of cancer-related deaths (32). Despite advances in new therapies, metastatic colorectal cancer remains a poor prognosis (33, 34). Consistently, the findings of this study demonstrated that distant metastases of colorectal cancer, including liver, lung, bone and brain, should be considered as predictive factors for cancer-specific early death in ECRC patients. Among these risk factors, brain metastases have the most serious effect on the outcome of ECRC patients. Brain metastases were often seen in the later stages disease. The median survival for patients who develop brain metastases from CRC is only 3–6 months (35, 36). Liver metastases, the most common site of metastases, occur in about 50 per cent of patients (37). With the development of treatment, in addition to surgical resection, liver metastases can be excised after receiving neoadjuvant chemotherapy, and can also be treated by radiofrequency ablation and selective intraoperative radiotherapy(32).
Tumor size is an independent prognostic variable that may influence early death in ECRC patients in this study. Previous studies have identified the size of the tumor as a significant predictor of prognosis in colorectal cancer patients, which is in line with the findings of the study (10, 14, 38). Tumor size as a predictor has been intensively studied and validated. However, the cut-off value at which tumor size affects prognosis has not been clearly defined or is inconsistent in these previous studies. This study chose an appropriate cut-off value, which was a tumor size ༞ 4.5 cm, to verify its positive association with early death in ECRC patients.
Many studies have identified race as a prognostic factor of colorectal cancer (14, 39). This study found that race may be a risk factor for early death in ECRC patients. In contrast, five-year survival rates are lower among blacks than among whites in the U.S. (40). Other studies have also concluded that the mortality risk is higher for blacks than for other races (41). We investigated relevant articles to explain this contradiction. We found that the average age at first diagnosis was higher for whites than for other races, at 72 years (42). Therefore, this may lead to implications for the investigation of older colorectal cancer patients, which could cause contrary results to previous studies.
Clinically, doctors usually use TNM staging to assess a patient's prognosis (43). However, research has also pointed out the deficiencies of the TNM staging, even that it can mislead clinicians about treatment options(9). This study affirmed the reliability of the TNM staging in assessing the prognosis of patients. Besides, we found that T stage and N stage were risk indicators for early death in patients with ECRC.
The first nomogram prediction model for early death in ECRC patients was developed in this study. Nomograms take into account more prognostic factors than the traditional TNM staging. Therefore, it is now widely used for prognostic prediction of cancer patients (11). In addition, one of the advantages of nomograms is that they can provide patients with individualised predictions of their risk of early death. More importantly, from the perspective of the calibration curve, AUC, IDI, DCA, and NRI, our nomograms clearly demonstrated better prediction accuracy and prognostic value than the TNM staging system. The ROC curve presented clearly that nomograms performed more specific and sensitive. DCA curves also showed that integrated nomograms contribute to better clinical decision-making in individual treatment. These were both supported by the results of the IDI and NRI.
Despite the strengths demonstrated by the nomograms in this study, there are some limitations that remain to be considered. First, there was a selection bias in our sample. As the SEER database has the disadvantage of incomplete demographic and clinical information, some cases may be excluded due to missing data, and this may affect the precision of the nomograms. As the database did not contain information on the histological type of the tumour and specific aspects of treatment (time of the treatment started, detailed treatment plan and and adequacy of treatment, etc.), this may also mean that the effect of the histological type of the tumour and the adequacy of treatment was not fully analysed. Second, the prognostic factors included in this study were limited, such as biomarker expression status, and others that may have an effect on prognosis were not covered. In addition, without patient reported outcome measures (PROMs) included in the database, patients' subjective assessments of their state and the quality of their lives cannot be analysed, which in turn does not provide patients with sufficient information to make informed decisions. Finally, this study lacks an external validation cohort. More prospective data and a combination of other factors are needed to validate this model further.