The effect of primary tumor radiotherapy in patients with Unresectable stage IV Rectal or Rectosigmoid Cancer: a propensity score matching analysis for survival

Background To evaluate the impact of primary tumor radiotherapy on survival in patients with unresectable metastatic rectal or rectosigmoid cancer. Methods From September 2008 to September 2017, 350 patients with unresectable metastatic rectal or rectosigmoid cancer were retrospectively reviewed in our center. All patients received at least 4 cycles of chemotherapy and were divided into two groups according to whether they received primary tumor radiotherapy. A total of 163 patients received primary tumor radiotherapy, and the median radiation dose was 56.69 Gy (50.4–60). Survival curves were estimated with the Kaplan–Meier method to roughly compare survival between the two groups. Subsequently, the 18-month survival rate was used as the outcome variable for this study. This study mainly evaluated the impact of primary tumor radiotherapy on the survival of these patients through a series of multivariate Cox regression analyses after propensity score matching (PSM). Results The median follow-up time was 21 months. All 350 patients received a median of 7 cycles of chemotherapy (range 4–12), and 163 (46.67%) patients received primary tumor radiotherapy for local symptoms. The Kaplan–Meier survival curves showed that the primary tumor radiotherapy group had a significant overall survival (OS) advantage compared to the group without radiotherapy (20.07 vs 17.33 months; P = 0.002). In this study, the multivariate Cox regression analysis after adjusting for covariates, multivariate Cox regression analysis after PSM, inverse probability of treatment weighting (IPTW) analysis and propensity score (PS)-adjusted model analysis consistently showed that primary tumor radiotherapy could effectively reduce the risk of death for these patients at 18 months (HR: 0.62, 95% CI 0.40–0.98; HR: 0.79, 95% CI: 0.93–1.45; HR: 0.70, 95% CI 0.55–0.99 and HR: 0.74, 95% CI: 0.59–0.94). Conclusion Compared with patients with stage IV rectal or rectosigmoid cancer who did not receive primary tumor radiotherapy, those who received primary tumor radiotherapy had a lower risk of death. The prescription dose (59.4 Gy/33 fractions or 60 Gy/30 fractions) of radiation for primary tumors might be considered not only to relieve symptoms improve the survival of patients with inoperable metastatic rectal or rectosigmoid cancer.


Background
From the cancer statistical data of 2019, the incidence rate of colorectal cancer was 39.42 per 100,000 population in the U.S. [1], and the rate was 27.47 per 100,000 people in China [2]. The proportion of rectal cancer among colorectal cancers was 49.7% in China, which is higher than the corresponding 30.4% in the U.S. Approximately 25% of colorectal cancer patients present have overt metastases, and an additional 25-35% of patients will develop metastases during the course of their disease [3]. Approximately 80-90% of patients with metastatic colorectal cancer were not able to undergo a radical surgery of metastatic lesions [3,4], therefore a surgical removal to the primary tumor was considered unnecessary, which leads to the fact that most of stage IV rectal cancer are always with primary lesion in their whole survival time. Benefitting from the combination therapies of chemotherapy and targeted drugs, unresectable stage IV colorectal cancer normally has a median survival time of 20.7-33.4 months, which has been reported by several classical studies [5][6][7][8]. At present, systemic chemotherapy is still the preferred treatment for stage IV unresectable colorectal cancer.
Radiotherapy or resection of the primary tumor is only recommended for patients with primary tumor progression by the National Comprehensive Cancer Network (NCCN) guidelines [9]. These patients usually have typical local symptoms such as obstruction, bleeding, and pain. Although there is some controversy, most of the studies show that resection of the primary tumor without metastasectomy not only relieves pelvic symptoms in patients but also improves their survival [10][11][12]. However, few studies have explored whether radiotherapy for primary tumors can also improve the survival of these patients. To provide more meaningful clinical evidence to answer this question, in this study, we retrospectively analyzed 350 patients with stage IV unresectable rectal or rectosigmoid cancer using propensity score matching (PSM) analyses to explore whether there were any survival benefits in patients who received primary tumor radiotherapy.

Methods
We retrospectively reviewed 366 patients who were initially diagnosed with stage IV unresectable rectal or rectosigmoid cancer from September 2008 to September   Table 1.
The follow-up period was defined as the time from the confirmed diagnosis of metastatic colorectal cancer until death or at least 18 months after the confirmed diagnosis. The survival time of the two groups was compared with the Kaplan-Meier method. The outcome variable was the 18-month survival rate, and whether the primary tumor was treated with radiation was used as the exposure variable. Covariates that may be related to outcome variables were screened for by referring to previous literature, clinical experience, and univariate Cox regression analysis. Multivariate Cox regression analysis was used to identify the independent effects of exposure variables on the outcome variable after adjusting for relevant covariates. This study was a retrospective observational study, not a randomized controlled trial (RCT), so selection bias was inevitable. To minimize the effect of bias, the propensity score matching (PSM) method can achieve a similar randomization effect, further verifying the previous analysis results. The matching algorithm used binary logistic regression, and the caliper value was set to 0.05. Given that PSM can cause sample loss, this study used inverse probability of treatment weighting (IPTW) as a sensitivity analysis to assess the stability of the results. Moreover, we further verified the results by adjusting the propensity score analysis (PS-adjusted

Results
The median follow-up time was 21 months. Patients who received primary tumor radiotherapy had more cycles of chemotherapy (35.58% vs 24.06%; P = 0.019) and were more likely to receive second-line chemotherapy (26.38% vs 16.58%; P = 0.014) than those who did not receive primary tumor radiotherapy. Patients with a lower Karnofsky performance status (KPS) score, moderate differentiation and T3 stage constituted a higher percentage in the primary tumor radiotherapy group. All of the other characteristics were similar between groups ( Table 1).
The Kaplan-Meier survival curves showed that the primary tumor radiotherapy group had a significant overall survival (OS) advantage compared to the group without radiotherapy (20.07 vs 17.33 months; P = 0.002; Fig. 2). The 18-month survival rates were 73.01 and 42.25%, respectively, for the groups with and without primary tumor radiotherapy.
Univariable Cox regression analysis was performed to assess the associations between covariates and the 18month survival rate, and the results are displayed in Table 2. The possible protective factors based on the univariable Cox regression analysis include older age, more chemotherapy cycles, second-line chemotherapy, better chemotherapy response, metastatic lesions treated with radiotherapy, and primary tumor treated with radiotherapy. The possible risk factors include a higher T stage, N stage, and M stage. Based on the univariate Cox regression analysis and the distribution of related factors in the two groups of patients, 10 variables (including age, KPS, T stage, N stage, M stage, differentiation, chemotherapy cycle, second-line chemotherapy, chemotherapy response, radiotherapy for metastatic tumor) were selected as the covariates that needed to be adjusted for in subsequent multivariate Cox regression analysis. After adjusting for the above covariates in the multivariate Cox regression analysis, the primary tumor radiotherapy group had a lower risk of death than the group without primary tumor radiotherapy (HR: 0.62, 95% CI 0.40-0.98; Table 4). The 10 variables were included in propensity score matching. There were 91 matched patients in each group after 1:1  individual matching without replacement. The matching situation is shown in Fig. 3, and the clinicopathological features are presented in Table 3. The primary tumor radiotherapy group still showed a lower risk of death than the group without primary tumor radiotherapy after propensity score matching (HR: 0.79, 95% CI: 0.93-1.45; Table 4). The sensitivity analyses using propensity scorebased IPTW and PS-adjusted models yielded similar results ( Table 4). The results from our sensitivity analyses were consistent with our primary analysis findings. Patients treated with primary tumor radiotherapy in this study had a lower risk of death than those treated without radiotherapy (HR: 0.70, 95% CI 0.55-0.99 and HR: 0.74, 95% CI: 0.59-0.94).

Discussion
With the application of oxaliplatin and irinotecan in combination with the fluorouracil regimen, the survival time of stage IV colorectal cancer ranged from 16 to 20 months [13][14][15]. After entering the era of targeted drugs combined with chemotherapy, the survival time of stage IV colorectal cancer has been significantly improved to 20.7-33.4 months.
For patients with stage IV colorectal cancer who cannot be cured by radical surgery, in general, resection or radiotherapy was used as a local treatment to relieve local obstruction, hemorrhage and pain. More clinical studies focus on the benefits of primary tumor resection alone. Although there are still controversies at present [16,17], most of the existing clinical studies show that resection of the primary tumor alone can not only reduce the incidence of local complications [18] but also seems to be beneficial in terms of patient survival [11,19,20]. However, there are limited data regarding the effect of primary tumor radiotherapy in stage IV unresectable rectal or rectosigmoid cancer, and most of these studies mainly observed the palliative effect [21][22][23][24]. To the best of our knowledge, very few studies have explored the effects of primary tumor radiotherapy on the survival of metastatic rectal cancer. For clinical researchers, the main reason is that there are many factors that can affect the survival of patients with stage IV rectal cancer, and there are large individual differences. In retrospective observational studies, conventional multivariate regression analysis has difficulty effectively removing interference of confounding factors and  selection bias from the results, which makes the analysis results lack reliability and consistency. Moreover, it is very difficult to implement such randomized controlled trials; for example, two previous trials (NCT01086618 and NCT01978249) were terminated due to recruitment problems. This study designed a series of analyses based on PSM to minimize interference from other confounding factors and selection bias on the research results.
In previous clinical studies on primary tumor radiotherapy for metastatic rectal cancer, the radiotherapy doses were generally low. Sager et al. reviewed many studies in which the radiotherapy dose delivered to the primary tumors ranged from 25 to 50 Gy [25]. When the α/β of the tumor was assumed to be 10 Gy for the biologically equivalent dose (BED), the BED of the above studies ranged from 37.5 to 53.1 Gy. In this study, the radiation dose of the primary tumor was significantly higher than that in previous clinical studies. Overall, 78% of patients completed the prescription dose (59.4 Gy in 33 fractions or 60 Gy in 30 fractions) of radiotherapy, the average radiation dose was 56.69 Gy, and the average BED was 67 Gy. Previous studies showed that a prescription dose of 54 Gy to 60 Gy (BED = 65 to 72 Gy) delivered to rectal tumors would achieve a significant tumor regression effect, and the percentage of patients with tumor regression grade (TRG) 1 and 2 was approximately 60 to 63.9% [26,27].
In this study, Kaplan-Meier survival analysis showed that the median survival times of the primary tumor groups with and without radiotherapy were 20.07 ± 8.98 months and 17.33 ± 7.34 months, respectively. This was consistent with previous studies on stage IV colorectal cancer patients who only received chemotherapy (median survival was 16 to 20 months), so we decided to use the 18-month survival rate as the outcome variable in this study. Furthermore, in this study, the priori selection of covariates was based on previous studies and the experience of the authors but also considered the results of the univariate analysis. Subsequently, multivariate Cox regression analysis after adjusting for covariates, analysis after PSM, IPTW analysis and PS-adjusted model analysis were performed to examine the reliability of the results. All analyses consistently showed that primary tumor radiotherapy could effectively reduce the risk of death for these patients at 18 months. According to the results of the different analysis models above, although the hazard ratio (HR) increased significantly, the reduction in the risk of death did not change, and the range of the confidence interval gradually narrowed. Our results became more conservative and accurate with the IPTW and PS-adjusted model analyses. A retrospective observational study similar to this study showed that palliative radiotherapy could improve the survival of patients with metastatic rectal cancer [28]. However, several deficiencies exist in the study. The study did not further analyze the location of the lesion (primary or metastatic) targeted by palliative radiotherapy, and it did not consider the dose. Chemotherapy, as an important factor affecting the survival of patients with metastatic rectal cancer, was not analyzed in this study. These deficiencies have been corrected in this study.
There are still some shortcomings and limitations in this study: (1) the time range of eligible patients included in this retrospective study was from September 2008 to September 2017. During this period, the price of bevacizumab and cetuximab in China were high, and these drugs were not covered by local health care insurance. Patients could rarely afford these medications, so this study did not select patients who received bevacizumab or cetuximab. The lack of targeted drugs will definitely reduce the survival benefit of patients and may affect the benefits of radiotherapy for primary tumors. (2) Compared with the 12 cycles recommended by the guidelines, the median number of chemotherapy cycles in this study was relatively low, at only 7 cycles. Fewer chemotherapy cycles will reduce the therapeutic efficacy for all patients and may have an uncertain impact on the benefits of primary tumor radiotherapy. (3) This study was a realworld study (observational clinical study). There might be some confounding factors outside of clinical cognition and previous literature reports that may affect the accuracy of the research results.

Conclusions
In this study, compared with patients with stage IV rectal or rectosigmoid cancer who did not receive primary tumor radiotherapy, those who received primary tumor radiotherapy had a reduced risk of death for 18 months. The dose pattern of 59.4 Gy in 33 fractions or 60 Gy in 30 fractions was acceptable during concurrent chemotherapy. These doses of radiation for primary tumors might be considered not only to relieve symptoms but also to improve the survival of patients with inoperable metastatic rectal or rectosigmoid cancer.