Adjuvant Chemotherapy for Stage II Right- and Left-Sided Colon Cancer: Analysis of SEER-Medicare Data

Abstract

Purpose

Survival benefit from adjuvant chemotherapy is established for stage III colon cancer; however, uncertainty exists for stage II patients. Tumor heterogeneity, specifically microsatellite instability (MSI) which is more common in right-sided cancers, may be the reason for this observation. We examined the relationship between adjuvant chemotherapy and overall 5-year mortality for stage II colon cancer by location (right- versus left-side) as a surrogate for MSI.

Methods

Using Surveillance, Epidemiology, and End Results (SEER)-Medicare data, we identified Medicare beneficiaries from 1992 to 2005 with AJCC stage II (n=23,578) and III (n=17,148) primary adenocarcinoma of the colon who underwent surgery for curative intent. Overall 5-year mortality was examined with Kaplan-Meier survival analysis and Cox proportional hazards regression with propensity score weighting.

Results

Eighteen percent (n=2,941) of stage II patients with right-sided cancer and 22% (n=1,693) with left-sided cancer received adjuvant chemotherapy. After adjustment, overall 5-year survival benefit from chemotherapy was observed only for stage III patients (right-sided: HR 0.64; 95% CI, 0.59–0.68, p<0.001 and left-sided: HR 0.61; 95% CI, 0.56–0.68, p<0.001). No survival benefit was observed for stage II patients with either right-sided (HR 0.97; 95% CI, 0.87–1.09, p=0.64) or left-sided cancer (HR 0.97; 95% CI, 0.84–1.12, p=0.68).

Conclusions

Among Medicare patients with stage II colon cancer, a substantial number receive adjuvant chemotherapy. Adjuvant chemotherapy did not improve overall 5-year survival for either right- or left-sided colon cancers. Our results reinforce existing guidelines and should be considered in treatment algorithms for older adults with stage II colon cancer.

Introduction

Adjuvant chemotherapy for patients with completely resected stage III colon cancer is considered standard clinical practice.¹ However, controversy exists surrounding the use of adjuvant chemotherapy for patients with stage II colon cancer. Despite evidence that adjuvant chemotherapy for stage II patients may not be beneficial and guidelines that do not routinely recommend its use for patients with stage II colon cancer,² a substantial number of patients are still receiving this treatment. Utilizing the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked database, Schrag et al.³ found that between 1991 and 1996, 27% of patients over age 65, with surgically resected stage II colon cancer received chemotherapy. This pattern was recently confirmed by O’Connor et al.⁴ utilizing an updated SEER-Medicare database from 1992 to 2005.

With recent attention on tumor biology and the recognition that there is a great deal of heterogeneity,^5–10 should we be asking whether there is a subset of patients with stage II colon cancer who would be more likely to benefit from adjuvant chemotherapy? Heterogeneity in microsatellite instability (MSI) status is common in colon cancers, and is an independent predictor of survival: MSI-high tumors have a better overall prognosis^11–15 and significantly decreased risk of metastases.¹⁴ Ribic et al.¹¹ investigated the effectiveness of adjuvant chemotherapy with fluorouracil for both stage II and stage III colon cancer patients by MSI status. They found a survival benefit from fluorouracil-based adjuvant chemotherapy for microsatellite-stable and MSI-low stage II and stage III colon cancers, but not for those with MSI-high tumors.

MSI is seen predominantly in right-sided colon cancers with an estimated 20–25% of right-sided stage II cancers being MSI-high; MSI-high tumors of the left colon rarely exist, across all stages.^{10,12,16–19} This raises the possibility for a differential benefit from chemotherapy based on tumor location. We used the linked SEER-Medicare dataset to examine the relationship between adjuvant chemotherapy and overall 5-year mortality for right- versus left-sided stage II colon cancer. We hypothesized that left-sided stage II cancers would demonstrate a greater benefit from receipt of adjuvant chemotherapy compared to right-sided cancers.

METHODS

This study was reviewed by the University of Wisconsin-Madison Health Sciences Institutional Review Board and determined to be exempt.

Data Sources

We utilized the linked SEER registry and Medicare claims databases to identify patients diagnosed with colon cancer between 1992 and 2005. SEER cancer registry information includes patient demographics, tumor characteristics, first course of treatment, and survival for persons newly diagnosed with cancer. The SEER-Medicare dataset has successfully linked 93% of individuals aged 65 or older at diagnosis to their Medicare record.^20,21

Patient Selection

All Medicare-enrolled patients aged 66 years and older diagnosed in a SEER area from 1992 to 2005 were eligible for our study if they had a diagnosis of colon (International Classification of Diseases for Oncology ICD-O-3 site codes 18.0-18.9, and 19.9) adenocarcinoma (ICD-O-3 morphology codes 8140-47, 8210-11, 8220-21, 8260-63, 8480-81, and 8490) that were diagnosed at either American Joint Committee on Cancer (AJCC) stage II or III. Patients with rectal cancer (site code 20.9) were excluded, as well as patients with mucinous cystadenocarcinoma (morphology code 8470).²² We further selected for patients who underwent surgery for likely curative intent within six months of diagnosis by using International Classification of Diseases, ninth revision, Clinical Modification (ICD-9-CM) procedure codes: 45.7x (partial excision of large intestine) and 45.8x (total intra-abdominal colectomy). Patients were also required to have continuous enrollment in Medicare Part A and B from 12 months prior to the date of their diagnosis to five years following date of discharge, death, or December 31, 2005 to facilitate ascertainment of overall health status, post-operative chemotherapy administration, and survival. Patients were excluded who were also enrolled in a Health Maintenance Organization during the same time period, diagnosed with another malignancy one year before or after the date of colon cancer diagnosis, or if their first diagnosis of colon cancer was made after death (i.e., on autopsy).

From an initial cohort of 46,627 patients, sequential exclusions were carried out for tumors with an unknown location or from the appendix (site codes 18.8, 18.9 and 18.1), unknown tumor grade, initial cancer treatment other than surgery (pre-operative radiation or chemotherapy), missing nodal assessment, and death within 30 days of surgery or prior to discharge from the surgical hospitalization; yielding a final sample size was 40,726 patients.

Outcome Variable

The primary outcome measure was all-cause five-year mortality, defined as death within 5 years of primary surgery for colon cancer from the SEER Patient Entitlement and Diagnosis Summary File (PEDSF).

Explanatory Variables

The primary explanatory variable was receipt of adjuvant chemotherapy. We defined postoperative chemotherapy based on prior work by Bradley et al,²³ Dobie et al,²⁴ and our recently published paper on adjuvant chemotherapy for stage II colon cancer patients with poor prognostic features.⁴ Claims were included within the following treatment window: beginning with the date of primary surgery and ending with (a) claim date after which there were 3 months free of colon cancer treatment; (b) evidence of distant cancer recurrence; or (c) 9 months after primary surgery, whichever came first.^23,24 Patients with at least one claim in this treatment window were classified as having received any chemotherapy, without differentiation between type of chemotherapy received. Capecitabine administration was captured using Healthcare Common Procedure Coding System (HCPCS) codes J8520 and J8521. Patients without any claims in this treatment window were considered to have received no chemotherapy.

Analyses were stratified by tumor location and AJCC stage. Dividing tumor location into right- and left-side follows a validated approach described by Meguid et al.²⁵ and Benedix et al.²⁶ and was used in our previously published paper examining the relationship between colon cancer location (right- versus left-side) and five-year mortality by stage.²⁷ Right-sided colon cancers were identified by ICD-O-3 site codes: 18.0–Cecum, 18.2–Ascending colon, 18.3–Hepatic flexure of colon, and 18.4–Transverse colon. Left-sided colon cancers were defined with codes 18.5–Splenic flexure of colon, 18.6–Descending colon, 18.7–Sigmoid colon, and 19.9–Rectosigmoid. AJCC stage II–III was identified from the SEER data.

Control Variables

Patient sociodemographic variables included age, gender, race/ethnicity, marital status, SEER registry region, and rural/urban county of residence. Census tract median level of household income and median level of education were used as proxies for patient socioeconomic status. Patient risk adjustment and treatment variables included hospitalization in the year prior to diagnosis, the presence of any in-hospital complication,²⁸ and a score for overall health status. Each patient was assigned a Centers for Medicare and Medicaid Services Hierarchical Condition Categories (HCC)²⁹ score. HCC score is a measure of healthcare utilization and was used as a marker of overall health status. Data on oncologist visit within 30 days of surgical discharge was also extracted. We also incorporated several cancer-specific variables, including year of diagnosis (divided into periods of colonoscopy coverage by Medicare)³⁰ and poor prognostic features (diagnosis in the setting of intestinal obstruction or perforation, emergent admission for surgery, T4 stage, poor/undifferentiated tumor histology, and fewer than 12 lymph nodes examined at surgery).²

Statistical Analyses

We compared the frequency of all patient sociodemographic, risk adjustment, treatment, and cancer-specific variables by receipt of chemotherapy (none vs. any) for right- and left-sided colon cancers within each stage (II–III) using χ² tests for categorical variables and two-way analysis of variance tests for continuous variables. Cox proportional hazards regression was conducted to obtain adjusted hazards ratios (HRs) and 95% confidence intervals (CIs) for all-cause five-year mortality for patients who did or did not receive chemotherapy by tumor location within each stage group, controlling for all other variables.

In clinical practice, decisions to treat patients with chemotherapy are based on a number of factors including age, comorbidities, tumor stage and grade, lymphovascular invasion, and patient/provider preferences. We further adjusted the Cox proportional hazard ratios with propensity score weighting to address potential bias in estimates of the true treatment effect given the observational nature of the dataset.^31–34 Propensity scores were calculated for each patient using models to estimate each patient’s probability of receiving adjuvant chemotherapy conditional on sociodemographic and clinical characteristics (Table 1). Propensity score weights based on the inverse of the propensity score were included in the fully adjusted models.³⁵ Kaplan-Meier survival analysis was adjusted for propensity scores. Overall five-year survival rates were compared for patients receiving any versus no adjuvant chemotherapy within each stage (II–III) and stratified by location (right- vs left-sided colon cancer).

Table 1.

Propensity score models

	Stage II Right-Sided Cancer	Stage II Left-Sided Cancer	Stage III (Right- and Left-Sided Cancer)
Main Effect Variables	Age	Age	Age
	Race	Race	Race
	In-hospital complications	Marital Status	Census-tract education
	Visit to oncologist within 30 days	In-hospital complications	In-hospital complications
	Tumor grade	Visit to oncologist within 30 days	Diagnosis in the setting of obstruction
		Tumor grade	Visit to oncologist within 30 days Tumor grade
Interactions	Age*Complication	Age*Complication	Age*Complication
	Age*Oncologist visit	Age*Oncologist visit	Age*Oncologist visit

Analyses were conducted with Stata 12.0 software (StataCorp, College Station, Texas). All tests of significance used two-sided p-values at the p<0.05 level; 95% CIs were calculated using robust estimates of the standard error.

RESULTS

Sample Characteristics

We included 40,726 patients in our study using the patient selection algorithm described in the Methods section and stratified groups by AJCC stage and tumor location. Eighteen percent (n=2,941) of stage II patients with right-sided cancer and 22% (n=1,693) with left-sided cancer received adjuvant chemotherapy compared to 56% (n=6,496) of stage III right-sided cancer and 60% (n=3,284) with stage III left-sided cancer.

Patients who received adjuvant chemotherapy were more likely to be younger, white, male, and married; have a higher household income; be diagnosed with a more advanced tumor; have a better HCC score; have less hospitalizations in the previous year; and be seen by an oncologist within 30 days of surgery. Similar patterns were seen within tumor location groups for stage II (Table 2) and stage III patients (Table 3).

Table 2.

Characteristics of Medicare beneficiaries undergoing resection for Stage II colon cancer stratified by location and chemotherapy (N=23,578)

	Right-Sided Cancer (n=15,992)			Left-Sided Cancer (n=7,586)
Characteristic	No Chemotherapy (n=13,051)	Any Chemotherapy (n=2,941)	p-value	No Chemotherapy (n=5,893)	Any Chemotherapy (n=1,693)	p-value
Patient Characteristics
Age, %			<0.001			<0.001
65–69	7.7	21.7		10.3	25.5
70–74	15.9	31.0		19.3	33.0
75–79	23.2	28.8		23.9	26.2
80–84	26.0	12.7		24.2	12.2
85+	27.2	5.8		22.4	3.1
Male gender, %	36.1	44.5	<0.001	44.8	50.6	<0.001
Race/ethnicity, %			0.016			0.024
White	87.7	87.2		83.9	85.9
Black	5.8	4.8		6.9	4.7
Asian or Pacific Islander	3.0	3.6		5.0	5.2
Hispanic or Other	3.5	4.3		4.2	4.2
Marital status, %			<0.001			<0.001
Married	44.6	60.2		48.0	62.4
Widowed	40.7	25.6		36.7	24.0
Single, Separated, or Divorced	11.6	11.1		12.1	11.0
Unknown	3.1	3.1		3.2	2.6
SEER registry, %			<0.001			0.023
Connecticut	11.3	7.8		11.3	9.1
Detroit	9.5	12.6		10.3	11.3
Hawaii	1.4	1.2		2.1	1.4
Iowa	14.0	13.3		12.3	13.2
New Mexico	2.3	2.0		3.1	2.8
Seattle	7.7	7.5		7.6	6.5
Utah	2.8	2.2		3.1	3.1
Atlanta & Rural Georgia	3.7	4.0		4.2	4.3
Kentucky	4.4	5.5		4.4	5.1
Louisiana	3.7	3.9		4.4	5.4
New Jersey	9.5	10.5		8.7	10.3
California	29.8	29.5		28.8	27.6
Residence location, %			0.003			0.121
Major metropolitan	55.2	58.7		56.0	56.7
Metropolitan or urban	34.5	31.4		34.0	32.0
Less urban or rural	10.3	10.0		10.0	11.3
Median household income (census tract), $ in thousands, mean (SD)	45.0 (21.9)	46.1 (22.7)	0.016	43.8 (22.3)	45.2 (22.1)	0.030
Less than 12 yrs education (census tract), %, mean (SD)	18.7 (12.2)	19.0 (12.6)	0.247	19.8 (12.9)	19.5 (13.0)	0.511
Year of diagnosis by Medicare colonoscopy coverage, %			<0.001			<0.001
1992–1997	33.0	31.4		38.2	34.9
1998-June 2001	24.2	28.0		23.3	27.9
July 2001–2005	42.9	40.6		38.5	37.2
Tumor Characteristics
Poor Prognostic Features
Emergent admission, %	18.9	15.8	<0.001	20.4	18.6	0.098
Intestinal obstruction on admission, %	3.3	3.3	0.957	7.2	8.8	0.026
Intestinal perforation on admission, %	1.4	2.3	<0.001	2.7	5.5	<0.001
T stage, %			<0.001			<0.001
T3	87.5	79.9		86.7	80.0
T4	12.6	20.1		13.3	20.0
Number of nodes resected, %			0.007			0.19
0–11 nodes	49.3	46.6		66.2	64.5
12 or more nodes	50.7	53.4		33.8	35.5
Tumor grade, %			<0.001			0.415
Well-differentiated	7.3	6.4		8.7	8.2
Moderately-differentiated	69.9	65.2		79.6	79.0
Poorly or undifferentiated	22.8	28.4		11.7	12.8
Comorbidity & Treatment Measures
HCC risk score, mean (SD)	2.00 (1.39)	1.74 (1.19)	<0.001	2.06 (1.39)	1.77 (1.16)	<0.001
Any hospitalization in the previous year, %	29.8	20.7	<0.001	26.3	18.7	<0.001
Rehospitalization within 30 days of discharge, %	11.6	8.7	<0.001	12.7	7.9	<0.001
Any in-hospital surgical complication, %	3.0	2.2	0.013	5.2	5.9	0.287
Oncologist visit within 30 days of surgery, %	21.9	55.3	<0.001	22.6	53.3	<0.001

AJCC = American Joint Committee on Cancer; SD = standard deviation; HCC = Hierarchical Condition Categories

Table 3.

Characteristics of Medicare beneficiaries undergoing resection for Stage III colon cancer stratified by location and chemotherapy (N=17,148)

	Right-Sided Cancer (n=11,655)			Left-Sided Cancer (n=5,493)
Characteristic	No Chemotherapy (n=5,159)	Any Chemotherapy (n=6,496)	p-value	No Chemotherapy (n=2,209)	Any Chemotherapy (n=3,284)	p-value
Patient Characteristics
Age, %			<0.001			<0.001
65–69	4.7	18.5		6.9	22.0
70–74	10.3	28.5		14.3	32.0
75–79	19.4	29.6		22.9	27.8
80–84	28.2	17.4		28.4	14.4
85+	37.3	6.0		27.5	3.7
Male gender, %	32.7	41.3	<0.001	44.5	52.2	<0.001
Race/ethnicity, %			0.001			0.012
White	85.3	85.6		79.9	81.9
Black	7.3	6.0		9.2	6.9
Asian or Pacific Islander	3.8	4.3		6.6	7.3
Hispanic or Other	3.6	4.2		4.4	3.9
Marital status, %			<0.001			<0.001
Married	38.4	58.2		43.9	61.9
Widowed	46.6	27.1		40.5	23.6
Single, Separated, or Divorced	12.0	11.8		12.5	11.9
Unknown	3.1	2.9		3.2	2.6
SEER registry, %			0.003			0.565
Connecticut	9.9	10.2		11.5	10.2
Detroit	10.1	11.3		10.5	10.1
Hawaii	1.8	1.6		2.7	2.9
Iowa	13.1	12.2		11.0	11.2
New Mexico	2.7	2.7		3.3	2.7
Seattle	7.7	7.2		6.2	6.4
Utah	3.4	3.2		3.8	3.4
Atlanta & Rural Georgia	3.9	4.0		4.0	3.9
Kentucky	4.7	5.0		4.9	4.7
Louisiana	3.6	4.4		3.9	5.1
New Jersey	9.9	11.6		8.9	9.9
California	29.4	26.5		29.2	29.3
Residence location, %			0.330			0.627
Major metropolitan	57.5	56.2		56.5	55.2
Metropolitan or urban	32.3	33.5		33.5	34.7
Less urban or rural	10.3	10.3		10.0	10.0
Median household income (census tract), $ in thousands, mean (SD)	43.8 (21.0)	46.4 (22.6)	<0.001	42.4 (20.2)	45.2 (22.3)	<0.001
Less than 12 yrs education (census tract), %, mean (SD)	19.3 (12.7)	18.5 (12.4)	0.0014	20.5 (13.1)	19.4 (12.7)	0.003
Year of diagnosis by Medicare colonoscopy coverage, %			<0.001			<0.001
1992–1997	33.1	28.2		36.4	31.4
1998-June 2001	23.5	25.7		21.5	25.0
July 2001–2005	43.4	46.1		42.1	43.6
Tumor Characteristics
Poor Prognostic Features
Emergent admission, %	25.9	17.1	<0.001	26.8	16.0	<0.001
Intestinal obstruction on admission, %	4.9	3.9	0.013	8.7	6.4	0.001
Intestinal perforation on admission, %	2.4	1.3	<0.001	3.1	1.7	0.001
T stage, %			<0.001			<0.001
T1, or T2	8.8	10.9		9.8	13.1
T3	70.4	70.3		71.3	70.7
T4	20.8	18.8		18.9	16.2
Number of nodes resected, %			<0.001			0.029
0–11 nodes	46.0	42.5		62.4	59.5
12 or more nodes	54.0	57.5		37.6	40.5
Tumor grade, %			0.158			0.944
Well-differentiated	4.3	4.7		5.8	5.6
Moderately-differentiated	58.9	60.1		73.2	73.5
Poorly or undifferentiated	36.8	35.2		21.1	20.9
Comorbidity & Treatment Measures
HCC risk score, mean (SD)	2.9 (1.6)	2.3 (1.2)	0.0014	2.9 (1.6)	2.3 (1.3)	<0.001
Any hospitalization in the previous year, %	33.1	22.9	<0.001	30.9	18.5	<0.001
Rehospitalization within 30days of discharge, %	16.0	8.6	<0.001	15.5	8.7	<0.001
Any in-hospital surgical complication, %	3.8	2.6	<0.001	6.5	3.6	<0.001
Oncologist visit within 30 days of surgery, %	27.5	59.2	<0.001	25.6	55.9	<0.001

AJCC = American Joint Committee on Cancer; SD = standard deviation; HCC = Hierarchical Condition Categories

Survival Benefit of Chemotherapy by Tumor Location

Unadjusted models showed a significant survival benefit for all groups from adjuvant chemotherapy for all groups: stage II right-sided (HR 0.66; 95% CI, 0.61 to 0.72), stage II left-sided (HR 0.61; 95% CI, 0.55 to 0.68), stage III right-sided (HR 0.50; 95% CI, 0.47 to 0.53), and stage III left-sided (HR 0.44; 95% CI, 0.40 to 0.47) (Table 4). However, after adjustment for all patient sociodemographic, risk adjustment, treatment, and cancer-specific variables, patients with stage II right-sided cancer no longer showed a benefit from adjuvant chemotherapy (HR 0.92; 95% CI, 0.85 to 1.01). Finally, when propensity score weighting was included to address selection bias owing to nonrandom treatment assignment, a survival benefit was no longer found for either group of stage II colon cancer patients (right-side: HR 0.97; 95% CI, 0.87 to 1.09; left-side: HR 0.97; 95% CI, 0.84 to 1.12).

Table 4.

Adjusted hazard ratios and 95% confidence intervals for 5 year mortality, by stage and location

	Stage II Right-Sided (n=15,992)			Stage II Left-Sided (n=7,586)			Stage III Right-Sided (n=11,655)			Stage III Left-Sided (n=5,493)
	HR	95% CI	p-value	HR	95% CI	p-value	HR	95% CI	p-value	HR	95% CI	p-value
Unadjusted
No Chemotherapy	1.00			1.00			1.00			1.00
Any Chemotherapy	0.66	0.61, 0.72	<0.001	0.61	0.55, 0.68	<0.001	0.50	0.47, 0.53	<0.001	0.44	0.40, 0.47	<0.001
Adjusted for all covariates a
No Chemotherapy	1.00			1.00			1.00			1.00
Any Chemotherapy	0.92	0.85,1.01	0.07	0.83	0.74, 0.93	0.001	0.64	0.60 0.68	<0.001	0.59	0.54, 0.65	<0.001
Adjusted for all covariates with propensity score weightingb
No Chemotherapy	1.00			1.00			1.00			1.00
Any Chemotherapy	0.97	0.87, 1.09	0.64	0.97	0.84, 1.12	0.68	0.64	0.59, 0.68	<0.001	0.61	0.56, 0.68	<0.001
HR, hazard ratio; CI, confidence interval

^aCox regression model controlling for age, gender, ethnicity, marital status, residence, SEER registry, year of diagnosis by Medicare colonoscopy coverage, census-tract income and education, hospitalizations in prior year, any in-hospital complication, HCC risk score, and poor prognostic features

^bSee Table 1 for variables included in the propensity score models

Kaplan-Meier survival analysis with propensity score weighting compared 5-year survival between patients who received any versus no adjuvant chemotherapy by tumor location (right- versus left-side) within each stage group (Fig 1a through d). No difference was observed between treated and untreated patients with either stage II right-sided (Wald χ²=0.27, p=0.60) or stage II left-sided cancers (Wald χ²=0.28, p=0.60) by Cox regression-based test for equality of survival curves. In contrast, all groups of stage III patients derived a significant survival benefit from chemotherapy (right-side: Wald χ²=243.91; p<0.001; left-side: Wald χ²=133.37; p<0.001).

Figure 1.

Kaplan-Meier survival curves showing overall survival by chemotherapy administration, tumor location (right- vs left-side), and stage group with propensity score weightin(Panel 1a: Stage II right-sided disease, Panel 1b: Stage II left-sided disease, Panel 1c: Stage III right-sided disease, Panel 1d: Stage III left-sided disease

DISCUSSION

Adjuvant chemotherapy is standard of care for patients with resected stage III colon cancer,¹ while the benefit for stage II patients remains unclear. Some studies have found no significant survival benefit for stage II patients who receive this treatment,^3,4 while others report a definite benefit of adjuvant chemotherapy over surgery alone for stage II patients.^36,37 However, the majority of evidence shows only a small absolute survival benefit, in the range of 2–4%.^38–40 This small benefit needs to be compared to the high likelihood of side effects from treatment. Given the previously identified relationship between MSI and response to chemotherapy, our primary aim was to determine if a subset of patients with stage II colon cancer, defined by tumor location as a surrogate for MSI status, could be identified that received a survival benefit from adjuvant chemotherapy. We found that, after controlling for multiple patient, risk adjustment, treatment, and cancer-specific variables in addition to propensity score weighting, there was no survival benefit for either stage II right-sided or stage II left-sided colon cancer patients who received any adjuvant chemotherapy compared to those who did not. Consistent with previous studies,^{4,37,39,41–46} we do show a significant survival benefit for stage III patients who receive adjuvant chemotherapy, regardless of tumor location.

Colon tumor heterogeneity has been described by many groups.^5–7 More specifically, different gene expressions have been described between tumors in the right and left colon, such as higher MSI in right-sided cancers and more chromosomal instability in left-sided cancers.^8–10 In a previous study, Ribic et al.¹¹ found a differential benefit from fluorouracil-based adjuvant chemotherapy for both stage II and stage III colon cancers with microsatellite-stable and MSI-low tumors compared to those with MSI-high tumors. And further analysis of patients included in the MOSAIC trial³⁷ points to a potential benefit of adding oxaliplatin to adjuvant 5-fluorouracil/leucovorin in patients with MSI-high stage III colon cancer.⁴⁷ MSI-high tumors are more prevalent on the right side of the colon^10,16–19 and even more so in stage II right-sided cancers.⁴⁸ MSI status of a tumor is not always readily available. In this study, we used tumor location (right- versus left-side) as a proxy for MSI status and did not find any significant benefit for traditional fluorouracil-based adjuvant chemotherapy for either stage II right-sided or stage II left-sided colon cancers.

There are limitations to this study. First, we examine only Medicare beneficiaries age 66 years and older at the time of diagnosis, which limits the applicability. Molecular etiologies of MSI in colon cancers are associated with age. MSI-high tumors in patients less than 50 years old are four times more likely to be due to a germline mutation in a DNA mismatch repair gene than in patients older than 50, who are four times more likely to have MSI due to MLH1 methylation.¹⁶ This study should be repeated in a broader population before applying our results to younger patients. Second, unmeasured factors such as patient preferences or provider practice patterns may play a role in patient outcome. We attempted to limit the impact of these unmeasured factors by adjusting our models with propensity score weighting for receipt of chemotherapy and including a clinically relevant patient, risk-adjustment, and treatment variables in our final models. The major limitation of this study is the fact that MSI status was not available in this data set, which prohibits direct testing of our hypothesis that MSI status impacts benefit from adjuvant chemotherapy in this population of patients. A related limitation is that due to the administrative claim data from Medicare, we were unable to specifically look at survival benefit by type of adjuvant chemotherapy. This limitation is especially relevant in light of the recent findings from further analysis of patients in the MOSAIC trial.^37,47 Further studies are needed to examine the effect of different types of adjuvant chemotherapy by colon cancer location and MSI status.

Despite these limitations, our findings add to the existing literature surrounding the controversy of adjuvant chemotherapy in stage II colon cancers. We know that substantial numbers of patients with stage II colon cancer continue to receive adjuvant chemotherapy despite current guidelines that do not recommend routine use.² Our results reinforce these guidelines and show that a simple division of stage II colon cancers as right- or left-sided does not identify a subgroup of patients that might receive benefit from adjuvant chemotherapy. When added to the existing literature, our results highlight the fact that the relationship between tumor biology, tumor location, and response to adjuvant chemotherapy is not straightforward. Future efforts to personalize treatment of colon cancers will likely involve direct gene expression testing for each individual. This practice will also continue to play an important role in shaping future clinical trials that examine adjuvant chemotherapy regimens for colon cancer.

SYNOPSIS.

Among Medicare patients with stage II colon cancer, adjuvant chemotherapy does not improve overall 5-year survival for either right- or left-sided colon cancers. This lack of benefit should be considered in treatment algorithms for older adults with colon cancer.