Primary tumor location and survival in colorectal cancer: A retrospective cohort study

Table 1 Patient characteristics before propensity score matching

	LPTL1			RPTL1
	Cetuximab (n = 164)	Bevacizumab (n = 679)	P value	Cetuximab (n = 84)	Bevacizumab (n = 385)	P value
Age at index date2, years			0.727			0.801
Median	61.0	61.0		66.5	67.0
Sex, n (%)			0.749			0.296
Female	59 (36.0)	256 (37.7)		36 (42.9)	192 (49.9)
Male	105 (64.0)	423 (62.3)		48 (57.1)	193 (50.1)
Race, n (%)			0.042			0.506
Black or African American	17 (10.4)	61 (9.0)		7 (8.3)	43 (11.2)
White	114 (69.5)	435 (64.1)		57 (67.9)	264 (68.6)
Other race	27 (16.5)	109 (16.1)		9 (10.7)	46 (11.9)
Missing	6 (3.7)	74 (10.9)		11 (13.1)	32 (8.3)
Region, n (%)			0.001			0.008
Northeast	38 (23.2)	144 (21.2)		20 (23.8)	73 (19.0)
Midwest	14 (8.5)	133 (19.6)		12 (14.3)	99 (25.7)
South	54 (32.9)	246 (36.2)		24 (28.6)	140 (36.4)
West	40 (24.4)	105 (15.5)		15 (17.9)	47 (12.2)
Unknown	18 (11.0)	51 (7.5)		13 (15.5)	26 (6.8)
Practice type, n (%)			0.079			0.052
Academic	2 (1.2)	31 (4.6)		0	17 (4.4)
Community	162 (98.8)	648 (95.4)		84 (100)	368 (95.6)
Site of disease, n (%)			1.000			NA
Colon	111 (67.7)	460 (67.7)		84 (100)	385 (100)
Rectum	53 (32.3)	219 (32.3)		0	0
Side of colon, n (%)			0.973			0.745
Left side	86 (52.4)	364 (53.6)		0	0
Rectosigmoid	18 (11.0)	79 (11.6)		0	0
Rectum	52 (31.7)	206 (30.3)		0	0
Splenic flexure	8 (4.9)	30 (4.4)		0	0
Right side	0	0		70 (83.3)	329 (85.5)
Transverse colon	0	0		14 (16.7)	56 (14.5)
Group stage at initial diagnosis, n (%)			0.199			0.001
Stage I	2 (1.2)	26 (3.8)		1 (1.2)	5 (1.3)
Stage II	13 (7.9)	73 (10.8)		5 (6.0)	40 (10.4)
Stage III	44 (26.8)	145 (21.4)		37 (44.0)	87 (22.6)
Stage IV	101 (61.6)	411 (60.5)		41 (48.8)	253 (65.7)
Unknown	4 (2.4)	24 (3.5)		0	0
ECOG PS at index date2, 3, n (%)			0.176			0.026
0	39 (23.8)	192 (28.3)		27 (32.1)	80 (20.8)
1	22 (13.4)	117 (17.2)		9 (10.7)	83 (21.6)
2+	5 (3.0)	29 (4.3)		3 (3.6)	26 (6.8)
Missing	98 (59.8)	341 (50.2)		45 (53.6)	196 (50.9)
Modified CCI at index date2, 4, n (%)			0.526			0.339
0	136 (82.9)	557 (82.0)		64 (76.2)	293 (76.1)
1	20 (12.2)	86 (12.7)		17 (20.2)	59 (15.3)
2	7 (4.3)	21 (3.1)		3 (3.6)	24 (6.2)
3+	1 (0.6)	15 (2.2)		0	9 (2.3)
NRAS tested, n (%)	51 (31.1)	281 (41.4)	0.020	28 (33.3)	160 (41.6)	0.204
NRAS status among those tested5, n (%)			0.112			0.373
Mutation negative	49 (96.1)	256 (91.1)		28 (100)	144 (90.0)
Mutation positive	0	19 (6.8)		0	14 (8.8)
Results pending	0	1 (0.4)		0	1 (0.6)
Unknown	1 (2.0)	1 (0.4)		0	0
Unsuccessful/indeterminate	1 (2.0)	4 (1.4)		0	1 (0.6)
BRAF tested, n (%)	46 (28.0)	313 (46.1)	< 0.001	35 (41.7)	176 (45.7)	0.579
BRAF status among those tested5, n (%)			0.075			0.072
Mutation negative	45 (97.8)	288 (92.0)		24 (68.6)	103 (58.5)
Mutation positive	0	23 (7.3)		10 (28.6)	73 (41.5)
Results pending	0	1 (0.3)		0	0
Unknown	1 (2.2)	1 (0.3)		0	0
Unsuccessful/indeterminate	0	0		1 (2.9)	0
First-line chemotherapy backbone, n (%)			< 0.001			< 0.001
FOLFIRI	105 (64.0)	165 (24.3)		64 (76.2)	96 (24.9)
FOLFOX	59 (36.0)	514 (75.7)		20 (23.8)	289 (75.1)
History of adjuvant chemotherapy6, n (%)			0.171			< 0.001
No/not documented	114 (69.5)	510 (75.1)		44 (52.4)	299 (77.7)
Yes	50 (30.5)	169 (24.9)		40 (47.6)	86 (22.3)
First-line therapy duration7, days			0.076			0.969
Median	197.0	183.0		156.0	162.0
Length of follow-up8, days			0.019			0.345
Median	501.5	442.0		323.5	350.0

¹Left-sided primary tumor location was defined as tumors that originated in the splenic flexure, descending colon, sigmoid colon, or rectum, and right-sided primary tumor location was defined as tumors that originated in the appendix, cecum, ascending colon, hepatic flexure, or transverse colon.

²Index date defined as the start date (first administration or non-cancelled order) of a first-line regimen containing cetuximab or bevacizumab.

³ECOG PS may have been recorded up to 30 d prior, or up to 7 d after, the index date, whichever was closest to the index date.

⁴CCI calculations exclude cancer diagnoses and include only comorbidities that were documented by the treating physician at any time prior to metastatic diagnosis date.

⁵Biomarker testing could occur at any point on, or up to, 28 d after the index date. For instances where multiple biomarker tests were available, the result from the successful test closest to the index date was used. Only the positive result among those tested is shown here; other categories included mutation negative, results pending, unknown, and unsuccessful/indeterminate test.

⁶Adjuvant therapy was only measured in non–stage IV patients.

⁷To account for potential censoring of patients on first-line therapy, the duration of first-line regimen was calculated using the following survival analysis methods: 1) the first administration or non-cancelled order for first-line therapy was the start date, 2) the last administration or non-cancelled order for first-line therapy was the end date, and 3) patients were classified as having discontinued first-line therapy (i.e. an event) if any of the following occurred: a) a patient started a subsequent line of therapy, b) a patient died, or c) a gap of more than 90 d occurred between a patient’s last administration or non-cancelled order for first-line therapy and the last activity date.

⁸Median follow-up time was calculated using observed time for all individuals, regardless of their outcome (i.e. not using survival analysis). Median was assessed using Kruskall-Wallis test. CCI: Charlson Comorbidity Index; ECOG PS: Eastern Cooperative Oncology Group performance status; FOLFIRI: 5-fluorouracil/leucovorin/irinotecan; FOLFOX: 5-fluorouracil/leucovorin/oxaliplatin; LPTL: Left-sided primary tumor location; NA: Not applicable; RPTL: Right-sided primary tumor location.

Table 2 Patient characteristics after propensity score matching

	Total (n = 792)	Cetuximab (n = 228)	Bevacizumab (n = 564)	P value
Age at index date1, years				0.923
Median	64.0	64.0	63.0
Sex, n (%)				0.914
Female	334 (42.2)	92 (40.4)	242 (42.9)
Male	458 (57.8)	136 (59.6)	322 (57.1)
Race, n (%)				0.905
Black or African American	71 (9.0)	22 (9.6)	49 (8.7)
White	549 (69.3)	158 (69.3)	391 (69.3)
Other race	111 (14.0)	32 (14.0)	79 (14.0)
Missing	61 (7.7)	16 (7.0)	45 (8.0)
Region, n (%)				0.620
Northeast	185 (23.4)	54 (23.7)	131 (23.2)
Midwest	113 (14.3)	26 (11.4)	87 (15.4)
South	253 (31.9)	75 (32.9)	178 (31.6)
West	158 (19.9)	48 (21.1)	110 (19.5)
Unknown	83 (10.5)	25 (11.0)	58 (10.3)
Practice type, n (%)				<0.001
Academic	35 (4.4)	0	35 (6.2)
Community	757 (95.6)	228 (100)	529 (93.8)
Site of disease, n (%)				1.000
Colon	630 (79.5)	180 (78.9)	450 (79.8)
Rectum	162 (20.5)	48 (21.1)	114 (20.2)
Group stage at initial diagnosis, n (%)				0.763
Stage I	9 (1.1)	3 (1.3)	6 (1.1)
Stage II	68 (8.6)	18 (7.9)	50 (8.9)
Stage III	245 (30.9)	79 (34.6)	166 (29.4)
Stage IV	458 (57.8)	124 (54.4)	334 (59.2)
Unknown	12 (1.5)	4 (1.8)	8 (1.4)
ECOG PS at index date1, 2, n (%)				0.379
0	197 (24.9)	61 (26.8)	136 (24.1)
1	130 (16.4)	30 (13.2)	100 (17.7)
2+	35 (4.4)	6 (2.6)	29 (5.1)
Missing	430 (54.3)	131 (57.5)	299 (53.0)
Modified CCI at index date1, 3, n (%)				0.861
0	650 (82.1)	185 (81.1)	465 (82.4)
1	104 (13.1)	32 (14.0)	72 (12.8)
2	34 (4.3)	10 (4.4)	24 (4.3)
3+	4 (0.5)	1 (0.4)	3 (0.5)
NRAS tested, n (%)	274 (34.6)	70 (30.7)	204 (36.2)	0.410
NRAS status among those tested4, n (%)				0.908
Mutation negative	268 (97.8)	68 (97.1)	200 (98.0)
Mutation positive	0	0	0
Results pending	1 (0.4)	0	1 (0.5)
Unknown	1 (0.4)	1 (1.4)	0
Unsuccessful/indeterminate	4 (1.5)	1 (1.4)	3 (1.5)
BRAF tested, n (%)	290 (36.6)	76 (33.3)	214 (37.9)	0.955
BRAF status among those tested4, n (%)				0.881
Mutation negative	248 (85.5)	64 (84.2)	184 (86.0)
Mutation positive	39 (13.4)	10 (13.2)	29 (13.6)
Results pending	1 (0.3)	0	1 (0.5)
Unknown	1 (0.3)	1 (1.3)	0
Unsuccessful/indeterminate	1 (0.3)	1 (1.3)	0
First-line chemotherapy backbone, n (%)				0.485
FOLFIRI	387 (48.9)	149 (65.4)	238 (42.2)
FOLFOX	405 (51.1)	79 (34.6)	326 (57.8)
History of adjuvant chemotherapy5, n (%)				0.789
No/not documented	519 (65.5)	140 (61.4)	379 (67.2)
Yes	273 (34.5)	88 (38.6)	185 (32.8)
First-line therapy duration6, days				0.424
Median	179.0	162.0	190.0
Length of follow-up7, days				0.533
Median	434.0	428.5	434.5
Side of colon, n (%)				0.868
Left side	283 (35.7)	78 (34.2)	205 (36.3)
Rectosigmoid	62 (7.8)	17 (7.5)	45 (8.0)
Rectum	154 (19.4)	47 (20.6)	107 (19.0)
Splenic flexure	19 (2.4)	5 (2.2)	14 (2.5)
Right side	231 (29.2)	68 (29.8)	163 (28.9)
Transverse colon	43 (5.4)	13 (5.7)	30 (5.3)

¹Index date defined as the start date (first administration or non-cancelled order) of a first-line regimen containing cetuximab or bevacizumab.

²ECOG PS may have been recorded up to 30 d prior, or up to 7 d after, the index date, whichever was closest to the index date.

³Charlson Comorbidity Index calculations exclude cancer diagnoses and include only comorbidities that were documented by the treating physician at any time prior to metastatic diagnosis date.

⁴Biomarker testing could occur at any point on, or up to, 28 d after the index date. For instances where multiple biomarker tests were available, the result from the successful test closest to the index date was used.

⁵Adjuvant therapy was only measured in non–stage IV patients.

⁶To account for potential censoring of patients on first-line therapy, the duration of first-line regimen was calculated using the following survival analysis methods: (1) The first administration or non-cancelled order for first-line therapy was the start date; (2) The last administration or non-cancelled order for first-line therapy was the end date; and (3) Patients were classified as having discontinued first-line therapy (i.e. an event) if any of the following occurred: (a) A patient started a subsequent line of therapy; (b) A patient died; or (c) A gap of more than 90 d occurred between a patient’s last administration or non-cancelled order for first-line therapy and the last activity date.

⁷Median follow-up time was calculated using observed time for all individuals, regardless of their outcome (i.e. not using survival analysis). Median was assessed using Kruskall-Wallis test. CCI: Charlson Comorbidity Index; ECOG PS: Eastern Cooperative Oncology Group performance status; FOLFIRI: 5-fluorouracil/leucovorin/irinotecan; FOLFOX: 5-fluorouracil/leucovorin/oxaliplatin.

Table 3 Cox regression estimation in the propensity score–matched sample (n = 792)

	Log (HR)	HR	SE [Log (HR)]	Robust SE [Log (HR)]	Z	P value
Cetuximab	0	1.00	0.18	0.19	0	0.996
LPTL	-0.58	0.56	0.13	0.15	-3.89	< 0.001
Cetuximab × LPTL	-0.14	0.87	0.24	0.25	-0.57	0.566

Data for right-sided primary tumor location were used as a reference in this model. HR: Hazard ratio; LPTL: Left-sided primary tumor location; SE: Standard error.

Table 4 Overall survival sensitivity analyses

	Matched/weighted	HR (95%CI) of cetuximab vs bevacizumab, for LPTL	HR (95%CI) of cetuximab vs bevacizumab, for RPTL
	mOS (mo) for cetuximab vs bevacizumab
Main analysis with propensity score matching	27.9 vs 25.4	0.87 (0.63-1.19)	1.00 (0.68-1.46)
Planned sensitivity analyses
1:1 matching without a caliper (propensity score matching)	27.9 vs 21.7	0.76 (0.55-1.06)	0.91 (0.59-1.41)
2:1 matching without a caliper (propensity score matching)	27.9 vs 22.3	0.82 (0.61-1.11)	0.87 (0.58-1.29)
IPTW	27.9 vs 25.6	0.87 (0.63-1.19)	0.98 (0.65-1.49)
LPTL only (propensity score matching)	29.4 vs 30.5	0.88 (0.61-1.28)	NA
RPTL only (propensity score matching)	17.9 vs 20.4	NA	1.09 (0.74-1.63)
Change definition of tumor location (IPTW)
RPTL: C/A/HF; LPTL: transverse to rectum	27.9 vs 25.6	0.89 (0.66-1.20)	0.96 (0.61-1.53)
RPTL: C/A/HF to transverse; LPTL: splenic flexure to rectosigmoid	27.9 vs 26.6	0.98 (0.66-1.44)	0.98 (0.64-1.50)
RPTL: C/A/HF to transverse; LPTL: splenic flexure to descending/sigmoid	27.4 vs 25.8	0.93 (0.61-1.43)	0.98 (0.65-1.48)
RPTL: C/A/HF; LPTL: transverse to rectosigmoid	27.9 vs 26.6	1.02 (0.71-1.46)	0.95 (0.60-1.52)
RPTL: C/A/HF; LPTL: transverse to descending/sigmoid	27.4 vs 25.8	0.98 (0.67-1.45)	0.95 (0.60-1.50)
RPTL: C/A/HF; LPTL: splenic flexure to rectum	28.3 vs 25.8	0.88 (0.64-1.21)	0.96 (0.60-1.51)
RPTL: C/A/HF; LPTL: splenic flexure to rectosigmoid	28.3 vs 26.6	1.00 (0.68-1.47)	0.93 (0.59-1.48)
RPTL: C/A/HF; LPTL: splenic flexure to descending/sigmoid	28.3 vs 26.0	0.95 (0.62-1.46)	0.94 (0.60-1.47)
NRAS mutation–negative patients only (IPTW)	28.9 vs 26.0	0.77 (0.40-1.50)	1.03 (0.54-1.97)
Restricted to patients with chemotherapy backbone of FOLFIRI (IPTW)	24.6 vs 23.1	0.96 (0.64-1.44)	1.05 (0.62-1.77)
Restricted to patients with chemotherapy backbone of FOLFOX (IPTW)	Not reached vs 25.91	0.63 (0.37-1.08)	0.52 (0.23-1.17)
Change definition of index date (propensity score matching)2	27.8 vs 24.8	0.86 (0.61-1.20)	0.96 (0.64-1.44)
Restrict to stage IV patients (IPTW)	29.7 vs 29.4	0.98 (0.61-1.57)	0.99 (0.51-1.90)

¹P < 0.05.

²Index date defined as the earliest administration/non-cancelled order date of cetuximab or bevacizumab within the index line of therapy. C/A/HF: Cecum/ascending/hepatic flexure; CI: Confidence interval; FOLFIRI: 5-fluorouracil/leucovorin/irinotecan; FOLFOX: 5-fluorouracil/ leucovorin/oxaliplatin; HR: Hazard ratio; IPTW: Inverse probability of treatment weighting; LPTL: Left-sided primary tumor location; mOS: Median overall survival; NA: Not applicable; RPTL: Right-sided primary tumor location.