According to the 2013 National Comprehensive Cancer Network guidelines, pelvic radiation therapy (RT) is one of the preferred regimens for patients with metastatic rectal cancer (mRC). The objective of this study was to analyze patterns of care and outcomes data for the receipt of RT among patients with mRC using the Surveillance, Epidemiology, and End Results (SEER) database.
Patients with stage IV rectal or rectosigmoid cancer were identified in the SEER database (2004-2009). Patients were stratified according to their primary disease site (rectum vs rectosigmoid), tumor (T) classification, and lymph node (N) classification. Treatment regimens (with or without surgical resection, with or without RT) were recorded. The Fisher exact test was used to compare RT rates based on stratified factors. Two and five-year survival rates were compared among treatment groups.
In total, 6873 patients with stage IV rectal cancer and 3417 patients with rectosigmoid cancer were identified. Overall, 20.5% of patients with rectal cancer underwent surgery alone, whereas 38.7% received RT with or without surgery. Within the rectosigmoid group, 51.4% of patients underwent surgery alone, and 15.1% received RT with or without surgery. The use of RT differed significantly between patients with in situ (Tis) through T2 tumors versus T3/T4 tumors (P < .001) and between those with N0 disease versus N1/N2 disease (P < .001). The 2-year and 5-year survival rates differed between treatment groups, with the highest survival rates observed among those who received combined surgery and RT.
Despite improvements in screening methods that allow for earlier detection and removal of precancerous polyps, colorectal cancer remains the third most common cancer in men and women in the United States, with an estimated 33% of these cases representing cancer within the rectum.[1, 2] Furthermore, of the projected 140,000 new diagnoses in 2013, approximately 20% to 25% of patients will present with metastatic disease and will have a projected median survival between 5 months and 6 months without treatment.[3, 4] Chemotherapy remains the primary treatment in the metastatic setting, and local treatments like surgery and radiotherapy (RT) play an important role in the control of pelvic disease. The National Comprehensive Cancer Network (NCCN) version 4 2013 guidelines for metastatic rectal cancer (mRC) stratify patients based on the presence of resectable or nonresectable metastases and include radiotherapy as a treatment option for both groups.
However, these treatment guidelines for mRC are highly variable and may include any combination of surgery (if resectable disease is present), RT, and/or chemotherapy. A few studies have sought to elucidate the role of pelvic RT—in either a neoadjuvant or adjuvant setting—for mRC; however, given the small patient cohorts and variable use of chemotherapy, the results were inconclusive.[6-8] In an effort to assess patterns of care for mRC in the United States, we conducted a study using the Surveillance, Epidemiology, and End Results (SEER) database to analyze the use of RT with or without surgery and to examine the survival outcomes associated with these treatment options.
MATERIALS AND METHODS
The SEER database was queried for cases between the years 1973 and 2009 that were coded as pertaining to either the “rectum,” with the primary site labeled as “C20.9-rectum-not otherwise specified (NOS),” or the “rectosigmoid junction,” with the primary site labeled as “C19.9-rectosigmoid junction,” and were categorized according to American Joint Committee on Cancer (AJCC) stage group “IV,” (stage IV: any tumor [T] classification [any T], any lymph node [N] classification [any N], and distant metastases [M1]). Separation according to AJCC stage grouping was available for patients from 1988 to 2009, with the AJCC Cancer Staging Manual sixth edition pertaining to patients from 2004 to 2009. A more detailed pattern of care and survival analysis was focused on diagnoses from 2004 to 2009 to reflect current clinical practices and data. These patients were stratified according to primary disease site (rectum vs rectosigmoid junction), T classification (T1-T4), and N classification (N1-N4). Then, data were collected with regard to initial treatment. Patients were classified as having received treatment according to one of the following treatment paradigms: 1) no RT and/or cancer-directed surgery; 2) external-beam RT alone; 3) external-beam RT before cancer-directed surgery; 4) external-beam RT after cancer-directed surgery; 5) radioactive implants; 6) RT, NOS; or 7) unknown. Only those patients who were denoted as undergoing their respective treatment regimens were recorded, and recommendations without follow-through for treatment were not included in calculations. The Fisher exact test was used to compare rates of RT use based on stratified factors.
Survival outcomes were again stratified according to primary tumor site (rectal vs rectosigmoid) as well as treatment regimens (with or without surgical resection, with or without RT, or a combination of both) and were calculated with SEER*Stat software version 7.1.10 (available at: www.seer.cancer.gov/seerstat; Accessed October 15, 2013) using the Kaplan-Meier method. Overall survival (OS) outcomes at 2 years and 5 years were selected for comparison.
From 1988 to 2009, a total of 10,681 patients with mRC and 6167 patients with metastatic rectosigmoid cancer (mRS) were identified in the SEER database. Subsequently, from 2004 to 2009, 6873 patients with mRC and 3417 patients with mRS were identified in the SEER database. Among patients in the mRC group, 20.5% initially underwent surgery alone, whereas 38.7% received RT with or without surgery (Table 1). Compared with patterns of care from 1988 to 2003, this finding reflects a significant change in treatment patterns, with increased use of RT for mRC over time, particularly from 2004 and onward (Fig. 1A). Among patients in the mRS group, 51.4% underwent surgery alone, and 15.1% received radiotherapy (with or without surgery) as part of their initial treatment (Table 1), consistent with the overall trend toward a surgery-alone approach (Fig. 1B). When assessing RT as a part of the treatment regimen, patients who had primary rectal tumors were significantly more likely to receive RT compared with those who had primary rectosigmoid tumors (P < .0001).
Table 1. Surgery and Radiotherapy Use for Metastatic Rectal and Rectosigmoid Cancer Between 2004 and 2009
No. of Patients (%)
Abbreviations: RT, radiotherapy.
This table provides a breakdown according to rectal versus rectosigmoid primary site for patients who were diagnosed from 2004 to 2009 and according to whether or not patients underwent surgery with or without RT. RT as a part of the treatment regimen was more common among patients with stage IV rectal primaries versus patients with rectosigmoid primaries (P < .0001).
RT with or without surgery
A detailed breakdown of the patterns of care according to T classification and N classification for both the mRC group and the mRS group from 2004 to 2009 is provided in Table 2. In the mRC group, stratification according to T classification revealed a significant increase in the use of RT among patients with earlier T classification (tumor in situ [Tis]-T2) versus later T classification (T3-T4): 71.2% of patients in the Tis through T2 group received RT compared with only 59.5% of patients in the T3 and T4 group (P < .001). A similar association between T classification was noted in the mRS group, in which patients who had an earlier T classification were more likely to receive RT versus those who had a later T classification (42.7% vs 20.3%; P < .0001). An assessment according to N classification revealed that patients in both groups who had lymph node-negative (N0) disease were significantly more likely to receive RT compared with patients who had lymph node-positive disease (mRC group, 74.7% [N0] vs 57% [N1/N2]; mRS group, 34.6% [N0] vs 10.5% [N1/N2]; P < .0001).
Table 2. Metastatic Rectal and Rectosigmoid Cancer Patterns of Care According to Tumor and Lymph Node Classification: 2004 to 2009a
No. of Patients
RT and Surgery
RT and Surgery
Abbreviations: RT, radiotherapy; Tis, tumor in situ.
The metastatic rectal and rectosigmoid cancer groups are stratified by treatment regimen as well as tumor and lymph node classification. The receipt of RT differed significantly between patients with Tis-T2 disease and those with T3-T4 disease (P < .001) and between patients with N0 disease versus those with N1-N2 disease (P < .001).
Lymph node classification
Survival outcomes at the 2-year and 5-year time points revealed a substantial difference in OS between treatment groups (Table 3). In the mRC group, surgery alone and RT alone were associated with 2-year OS rates of 45% and 28%, respectively. By comparison, the combined-modality (surgery plus RT) treatment groups had 2-year OS rates of 67% for those who received neoadjuvant RT, 54% for those who received adjuvant RT, and 63% for those who received neoadjuvant plus adjuvant RT. An assessment of 5-year OS outcomes demonstrated poorer survival outcomes in the monotherapy groups, including increased mortality rates with longer follow-up (Figs. 2A, 3A). For patients with mRC who underwent surgical monotherapy, only 28.9% of those who remained alive at 2 years were still alive at 5 years. Patients who received RT monotherapy had a similar late mortality, with a 5-year OS rate of 6%; and only 21.4% of those who remained alive at 2 years were still alive at 5 years. In contrast, the patients who received combined-modality treatment had 5- year OS rates of 24% to 28%; and from 40% to 45% of those who remained alive at 2 years were still alive at 5 years.
Table 3. Survival Outcomes Based on Treatment From 2004 to 2009a
Survival outcomes for patients with rectal and rectosigmoid primaries based on treatment regimen are displayed from 2004 to 2009. The 2-year and 5-year OS rates differed substantially between treatment groups, with the highest survival rates among those who received both RT and surgery.
RT and surgery
Neoadjuvant and adjuvant RT
For patients who received monotherapy, the 2-year and 5-year OS rates were similar in the mRC and mRS groups. Table 3 indicates that patients in both groups who received combined-modality treatment had similar 2-year OS rates (range, 54%-67%). The 5-year OS rates were higher for patients with mRC than for those with mRS (neoadjuvant RT: 28% vs 18%, respectively; adjuvant RT: 24% vs 9%, respectively; neoadjuvant plus adjuvant RT: 28% vs 0%, respectively).
A comparison of survival outcomes in the mRC group from 2004 to 2009 versus from 1988 to 2003 demonstrate similar trends in survival patterns according to treatment modality (Figs. 2, 3). Survival comparisons between earlier and more recent groups demonstrated improved OS in both the 2004 to 2009 mRC and mRS groups regardless of treatment modality (Tables 3, 4).
Table 4. Survival Outcomes Based on Treatment From 1988 to 2003
Survival outcomes for patients with rectal and rectosigmoid primaries based on treatment regimen are displayed from 1988 to 2003. The 2-year and 5-year OS rates differed substantially between treatment groups, with the highest survival rates among those who received both RT and surgery.
RT and surgery
Neoadjuvant and adjuvant RT
Currently, there is a great deal of flexibility in the NCCN guidelines concerning the use of RT for the treatment of mRC and mRS. This is driven by the variable disease presentation, which may include locally advanced disease and/or multiple sites of metastasis. However, patients who have limited disease burden may be cured or may live for many years with controlled or slowly progressive, distal disease. In such patients, control of both primary and metastatic disease is important to achieve either a cure or long-term symptom-free survival. Goals of intervention, thus, may be directed toward either a palliative or curative intent. The role of RT in achieving these goals is not clearly defined.
Multiple surgical series have demonstrated that long-term disease-free survival is achieved in approximately 20% of patients who present initially with mRC. However, debate continues regarding patients with mRC who have synchronous liver metastases and whether to treat the primary tumor first, second, or synchronously.[3, 9-11] This debate centers on the growing belief that it is the resectable metastatic lesions, and not the rectal primary, that dictate OS outcomes. Thus, once control of disseminated disease is obtained, curative treatment for the primary disease can be similar to that for nonmetastatic disease and may be performed using a combination of neoadjuvant and/or adjuvant chemoradiation followed by surgical resection.[3, 11] The current standard curative treatment for non-mRC comprises neoadjuvant chemoradiation followed by total mesorectal excisional surgery and adjuvant chemotherapy, as determined by the German Rectal Cancer Study Group Trial. Thus, patients with mRC who are treated with curative intent would receive standard therapy comprising combined surgery and chemoradiation.
Patterns of care indicate that clinicians demonstrate significant preferences regarding the use of RT in the treatment of mRC. The use of RT in this clinical scenario gained particular strength in 2004, coinciding with publication of the German Rectal Cancer Study Group trial by Sauer et al in 2004. The publication of that trial, which examined neoadjuvant versus adjuvant chemoradiation, also coincided with an observed increase in popularity of neoadjuvant treatment. Although the German Rectal Cancer Study Group trial may help explain the shift toward neoadjuvant RT with a concurrent decrease in adjuvant RT, it did not clearly account for the dominating increase in RT use alone within the mRC group (Fig. 1A). An explanation for why isolated RT use became a dominant treatment modality is unclear from that analysis, but improving systemic therapy may be a factor.
When the patterns of care over the past decade (2004-2009) are examined more closely, the current SEER database analysis establishes that RT is used more commonly as part of initial treatment for patients who have mRC compared with those who have mRS, in patients who have no clear lymph node involvement, and in patients who have lower T-classification tumors. Such treatment preferences likely reflect important considerations, such as tumor location, the ability to spare surrounding structures, and differences in lymphatic drainage and/or symptomatology. In the current analysis, patients with mRC were more likely to receive RT than patients with mRS, reflecting the known higher risk of pelvic failure associated with rectal cancer compared with rectosigmoid cancer. The increased use of RT among patients with Tis through T2 tumors and N0 disease may manifest the selection of patients who are more likely to be selected for a curative approach. Similarly, those with less disease burden (early T classification and early N classification disease) have a longer life expectancy and are more likely to be negatively impacted by a symptomatic pelvic recurrence during the course of their lifetime. Thus, they are more likely to receive pelvic RT to prevent such recurrences.
In addition, the propensity to use RT in the lower T-classification and N-classification groups may reflect the desire to reduce future palliative surgical intervention of large, bulky disease. It has been demonstrated that RT effectively allows for the control of pelvic symptoms while avoiding colostomy in patients with mRC. In the current series, patients with lower T-classification tumors (T0-T1) were more likely to receive RT monotherapy. This may be particularly true in patients who receive upfront treatment with chemotherapy; unfortunately, information regarding chemotherapeutic treatments was not available for correlation with our results (this is one of the limitations of the SEER database). It is important to note that palliative control can be obtained without surgical intervention. A study of 233 patients who received combination chemotherapy for the initial management of mRC indicated that 93% of patients did not require surgical intervention of their rectal tumors for palliative reasons. However, no difference in median survival was observed in the groups that did or did not receive emergent intervention (RT or surgery), indicating that upfront chemotherapy may improve symptomatology, but not necessarily OS.
There are little published data regarding RT administration to the primary tumor, and there are much less data that are independent of concurrent chemotherapy use. Two studies from Korea that attempted to address the question of adjuvant pelvic RT for stage IV rectal patients produced conflicting results, with 1 study reporting a decreased local failure rate and unknown OS outcomes and a second study reporting no improvements in local failure, disease-free survival, or OS. A study by Crane and colleagues suggested that chemoradiation with a biologic equivalent dose of > 35 grays in conjunction with surgery led to improved 2-year survival outcomes and symptomatic pelvic control compared with an RT-only group. Likewise, in the current study, we observed an association between initial RT treatment (especially in combination with surgery) and improved OS. In this SEER data set, patients who received neoadjuvant RT demonstrated the longest median survival. This finding was consistent when we examined OS results based on current patterns of care and when we evaluated survival results from previous treatment regimens (Figs. 2, 3). It should be noted that, although neoadjuvant RT demonstrated the greatest OS benefit regardless of the time period chosen, OS improved for each treatment modality when assessing current versus past outcomes—likely reflecting improvements in RT and surgical techniques along with the development of newer chemotherapeutics (Tables 3, 4).
Several important limitations exist when using the SEER database for analysis. The first limitation is the assumption that SEER coding is accurate, because it cannot be otherwise verified. In addition, SEER reporting does not specify whether RT was delivered to the primary site, the pelvis, metastatic sites, or a combination thereof. Given the ambiguity regarding RT site administration in the SEER records, an association between the use of RT and improved OS may reflect treatment not only to the primary tumor but also to metastatic lesions. However, because treatment modalities coded in the SEER database reflect the primary treatments given at the time of initial diagnosis, it may be reasonable to assume that a substantial portion of the coded RT and surgery treatments reflect management of the primary tumor, which is historically the focus of the first phase of treatment. Results obtained in this SEER analysis span many decades, during which there have been substantial improvements in surgical, medical, and RT techniques, such that patterns of care probably were strongly influenced by the era of treatment. An important future direction of studies similar to these will be to elucidate the role not only of treatment modality but also of the timing with which primary lesions versus metastatic lesions are treated.
Treatment recommendations for patients with mRC and mRS potentially may include surgery, RT, and/or chemotherapy. The variable nature of the rectal cancer disease burden makes it difficult to draw conclusions about patterns of care for mRC in the United States. From our analysis of SEER records, it appears that the inclusion of RT versus surgery alone is more common in patients with rectal primary tumors (vs rectosigmoid primary tumors), no lymph node involvement, or lower T-classification disease. Furthermore, treatment regimens that include both RT and surgery demonstrate prolonged survival, particularly in patients with rectal disease.