Cancer Chemother Pharmacol (2014) 73:409–416 DOI 10.1007/s00280-013-2367-7
ORIGINAL ARTICLE
CA19‑9 level as a prognostic and predictive factor of bevacizumab efficacy in metastatic colorectal cancer patients undergoing oxaliplatin‑based chemotherapy Yukiya Narita · Hiroya Taniguchi · Azusa Komori · Sohei Nitta · Kazuhisa Yamaguchi · Chihiro Kondo · Motoo Nomura · Shigenori Kadowaki · Daisuke Takahari · Takashi Ura · Masashi Andoh · Kei Muro
Received: 31 July 2013 / Accepted: 2 December 2013 / Published online: 10 December 2013 © Springer-Verlag Berlin Heidelberg 2013
Abstract Purpose The prognostic and predictive values of carbohydrate antigen 19-9 (CA19-9) levels in metastatic colorectal cancer (mCRC) remain unclear. We reviewed all mCRC patients at a single institution to evaluate the relationship between CA19-9 levels and survival. Methods Two hundred and fifty-two patients underwent first-line chemotherapy using oxaliplatin-based regimens between April 2005 and December 2009. The relationship between baseline CA19-9 levels and survival was analyzed. Moreover, we evaluated the relationship between baseline CA19-9 levels and clinicopathological factors. Results One hundred and fifty patients had elevated baseline CA19-9 levels (elevated group), and 79 patients had normal baseline CA19-9 (normal group) levels. Both KRAS and BRAF mutation rates were higher in the elevated group than in the normal group. Elevated CA19-9 level was a poor prognostic factor compared with normal CA19-9 levels (P = 0.0021). In the elevated group, the median survival time with bevacizumab was significantly longer with bevacizumab than without it (median OS, 27.8 vs. 15.3 months, P = 0.0019). However, the median survival time was not different with or without bevacizumab in the normal group (median OS, 36.5 vs. 38.0 months, P = 0.9515).
Electronic supplementary material The online version of this article (doi:10.1007/s00280-013-2367-7) contains supplementary material, which is available to authorized users. Y. Narita (*) · H. Taniguchi · A. Komori · S. Nitta · K. Yamaguchi · C. Kondo · M. Nomura · S. Kadowaki · D. Takahari · T. Ura · M. Andoh · K. Muro Department of Clinical Oncology, Aichi Cancer Center Hospital, 1‑1 Kanokoden, Chikusa‑ku, Nagoya, Aichi 464‑8681, Japan e-mail: yukiya.narita@aichi‑cc.jp
Conclusion Our results suggest that baseline CA19-9 level is an independent prognostic factor in mCRC patients, and it correlated with the KRAS/BRAF mutation status. Bevacizumab exhibits clinical activity only for high CA19-9 levels in mCRC. Keywords Bevacizumab · Carbohydrate antigen 19-9 · Colorectal cancer · Predictive factor · Prognostic factor
Background Colorectal cancer is the fourth leading cause of cancerrelated deaths in both sexes, causing an estimated 608,000 deaths worldwide in 2008 [2]. The 5-year overall survival rate of metastatic colorectal cancer (mCRC) patients remains low [2, 10]. Several prognostic factors for mCRC patients were found in previous studies. Köhne et al. [11] identified three risk groups of mCRC patients receiving 5-fluorouracil (5-FU)-based chemotherapy and developed a prognostic model centered on four baseline parameters: performance status, white blood cell count (WBC), alkaline phosphatase (ALP) levels, and number of metastatic sites. The addition of bevacizumab, an antibody drug that binds vascular endothelial growth factor (VEGF), to chemotherapy regimens improved mCRC patient outcomes in first-line treatments [5, 17]. Some studies suggest that the efficacy of bevacizumab does not depend on KRAS or BRAF mutation status [9, 16], although VEGF-D levels in tumor tissue may predict the effectiveness of bevacizumab treatment in combination with chemotherapy [23]. However, the value of VEGF-targeted therapy for patients with mCRC has not been established. Carbohydrate antigen 19-9 (CA19-9) [21] was first isolated from a colorectal cancer cell line, and it is a standard tumor marker for pancreatic cancer [8]. Although CA19-9
13
410
is overexpressed in 39 % mCRC patients [7], its levels alone are insufficient for monitoring colorectal cancer treatment according to American Society of Clinical Oncology guidelines [13]. CA19-9 acts as an adhesion factor for cells lining the blood vessels [21]. Furthermore, the association between CA19-9 levels and the efficacy of bevacizumab treatment remains unclear. The aim of this study was to determine the usefulness of pretreatment CA19-9 levels for predicting the survival and bevacizumab treatment efficacy in mCRC patients. We analyzed (1) the prognostic impact of baseline CA19-9 levels for mCRC patients receiving oxaliplatin-based first-line chemotherapy and (2) their ability to predict bevacizumab treatment efficacy.
Patients and methods
Cancer Chemother Pharmacol (2014) 73:409–416
serum CA19-9 level, KRAS/BRAF mutation status in tumor tissue, survival status, and date of last follow-up or death. To detect point mutations in KRAS codons 12 and 13, as well as the V600E mutation in the BRAF gene, we used the cycleave PCR technique [24]. The PCR reactions were performed using a cycleave PCR core kit (TAKARA, Co. Ltd., Ohtsu, Japan). Fluorescent signals were quantified using the Smart Cycler system (SC–100; Cepheid, Sunnyvale, CA, USA). Treatments The details of the treatment schedule are presented in Supplementary Data. Patients received a FOLFOX-4, modified FOLFOX-6, or CapeOX chemotherapy regimen [1, 5, 17]. Further, patients treated with bevacizumab [5, 17], cetuximab [14], and cediranib (AZD2171) [18] were included in this study. Bevacizumab has been approved for mCRC since April 2007 in Japan.
Patients Statistical analyses This study was a retrospective cohort study of mCRC patients who received oxaliplatin-based regimens (including FOLFOX-4, modified FOLFOX-6, and CapeOX) as first-line chemotherapy at Aichi Cancer Center Hospital. Patients were aged >20 years, had histologically proven inoperable colorectal cancer, and had serum CA19-9 levels available within 4 weeks of starting chemotherapy. Patients were excluded if they had double cancer or had previously undergone chemotherapy for advanced disease. CA19‑9 level measurement CA19-9 levels were measured at a certified laboratory in our institution using LUMIPULSE f® (Fujirebio, Inc.,) [15]. Reference range for normal CA19-9 levels was ≤37 U/ml [15]. Red cell phenotyping for Lewis antigen status was not performed in majority of the patients; therefore, patients, whose CA19-9 levels were persistently 70 years); gender (female vs. male); ECOG PS (0, 1 vs. 2, 3); primary tumor site (colon vs. rectum); number of metastatic sites (1 vs. ≥2); liver involvement (no vs. yes); prior colectomy (no vs. yes); histological subtypes (wel/mod vs. por/sig/muc); WBC count (50 ng/ml); and CA19-9 levels (>1, ≤37 vs. >37 IU/ml). A multivariate analysis using Cox proportional hazards model was performed to assess the relationship between clinicopathological variables and survival. To assess whether baseline CA19-9 is a prognostic marker independent of bevacizumab treatment and whether it predicts the patients’ response to bevacizumab, we separated patients into four groups on the basis of whether they underwent bevacizumab treatment and whether baseline CA19-9 levels were normal or elevated. We excluded the patients treated with cetuximab/cediranib from this analysis. We also assessed the relationship between baseline CEA and OS.
411
Cancer Chemother Pharmacol (2014) 73:409–416 Table 1 continued
Table 1 Patient characteristics Characteristics
N = 252
Characteristics
Age, years Median (range) Gender Female Male ECOG PS 0 1 2 3 Disease status Unresectable Recurrent Primary site Colon Rectum Number of metastatic sites 1 ≥2 Metastatic sites Liver Lung Lymph node Peritoneum Prior colectomy Yes No Histological subtypes Wel Mod Por Muc Sig Chemotherapy First-line FOLFOX/CapeOXa FOLFOX/CapeOXa + bevacizumab FOLFOX/CapeOXa + cetuximab FOLFOX/CapeOXa ± AZD2171 Second-line Third-line WBC (/μl) 37 IU/ml) Undetectableb (≤1 IU/ml)
143 (56.7) 109 (43.3) 159 (63.1) 93 (26.9) 79 (31.3) 150 (59.5) 23 (9.2)
KRAS/BRAF status 124 (49.2) 128 (50.8)
KRAS mutant G12X G13X
148 (58.7) 104 (41.3)
BRAF mutant V600E
116 (46.0) 136 (54.0) 118 (46.8) 146 (57.9) 153 (60.7) 193 (76.6) 210 (83.3) 42 (16.7) 55 (21.8) 166 (65.9) 22 (8.7) 5 (2.0) 4 (1.6)
133 (52.8) 101 (40.0) 8 (3.2) 10 (4.0) 190 (75.4) 99 (39.3) 237 (94.0) 15 (6.0) 128 (50.8) 124 (49.2)
KRAS wild type/BRAF wild type NA Outcome Median follow-up [range (months)] 30-day mortality Death Censored
34 (13.5) 10 (4.0) 7 (2.8) 76 (30.5) 124 (49.2) 23.3 (1.4–86.5) 0 (0) 197 (78.2) 55 (21.8)
ALP alkaline phosphatase, CA19-9 carbohydrate antigen 19-9, CEA carcinoembryonic antigen, ECOG Eastern Cooperative Oncology Group, LDH lactate dehydrogenase, Mod moderately differentiated adenocarcinoma, Muc mucinous carcinoma, NA not available, Por poorly differentiated adenocarcinoma, PS performance status, Sig signet ring cell carcinoma, WBC white blood cell count, Wel welldifferentiated adenocarcinoma a
FOLFOX: FOLFOX-4 or modified FOLFOX-6 (fluorouracil + leucovorin + oxaliplatin) b
Undetectable: Lewis antigen negative (Lea−b−)
Results were considered statistically significant when the two-sided P value was
ORIGINAL ARTICLE
CA19‑9 level as a prognostic and predictive factor of bevacizumab efficacy in metastatic colorectal cancer patients undergoing oxaliplatin‑based chemotherapy Yukiya Narita · Hiroya Taniguchi · Azusa Komori · Sohei Nitta · Kazuhisa Yamaguchi · Chihiro Kondo · Motoo Nomura · Shigenori Kadowaki · Daisuke Takahari · Takashi Ura · Masashi Andoh · Kei Muro
Received: 31 July 2013 / Accepted: 2 December 2013 / Published online: 10 December 2013 © Springer-Verlag Berlin Heidelberg 2013
Abstract Purpose The prognostic and predictive values of carbohydrate antigen 19-9 (CA19-9) levels in metastatic colorectal cancer (mCRC) remain unclear. We reviewed all mCRC patients at a single institution to evaluate the relationship between CA19-9 levels and survival. Methods Two hundred and fifty-two patients underwent first-line chemotherapy using oxaliplatin-based regimens between April 2005 and December 2009. The relationship between baseline CA19-9 levels and survival was analyzed. Moreover, we evaluated the relationship between baseline CA19-9 levels and clinicopathological factors. Results One hundred and fifty patients had elevated baseline CA19-9 levels (elevated group), and 79 patients had normal baseline CA19-9 (normal group) levels. Both KRAS and BRAF mutation rates were higher in the elevated group than in the normal group. Elevated CA19-9 level was a poor prognostic factor compared with normal CA19-9 levels (P = 0.0021). In the elevated group, the median survival time with bevacizumab was significantly longer with bevacizumab than without it (median OS, 27.8 vs. 15.3 months, P = 0.0019). However, the median survival time was not different with or without bevacizumab in the normal group (median OS, 36.5 vs. 38.0 months, P = 0.9515).
Electronic supplementary material The online version of this article (doi:10.1007/s00280-013-2367-7) contains supplementary material, which is available to authorized users. Y. Narita (*) · H. Taniguchi · A. Komori · S. Nitta · K. Yamaguchi · C. Kondo · M. Nomura · S. Kadowaki · D. Takahari · T. Ura · M. Andoh · K. Muro Department of Clinical Oncology, Aichi Cancer Center Hospital, 1‑1 Kanokoden, Chikusa‑ku, Nagoya, Aichi 464‑8681, Japan e-mail: yukiya.narita@aichi‑cc.jp
Conclusion Our results suggest that baseline CA19-9 level is an independent prognostic factor in mCRC patients, and it correlated with the KRAS/BRAF mutation status. Bevacizumab exhibits clinical activity only for high CA19-9 levels in mCRC. Keywords Bevacizumab · Carbohydrate antigen 19-9 · Colorectal cancer · Predictive factor · Prognostic factor
Background Colorectal cancer is the fourth leading cause of cancerrelated deaths in both sexes, causing an estimated 608,000 deaths worldwide in 2008 [2]. The 5-year overall survival rate of metastatic colorectal cancer (mCRC) patients remains low [2, 10]. Several prognostic factors for mCRC patients were found in previous studies. Köhne et al. [11] identified three risk groups of mCRC patients receiving 5-fluorouracil (5-FU)-based chemotherapy and developed a prognostic model centered on four baseline parameters: performance status, white blood cell count (WBC), alkaline phosphatase (ALP) levels, and number of metastatic sites. The addition of bevacizumab, an antibody drug that binds vascular endothelial growth factor (VEGF), to chemotherapy regimens improved mCRC patient outcomes in first-line treatments [5, 17]. Some studies suggest that the efficacy of bevacizumab does not depend on KRAS or BRAF mutation status [9, 16], although VEGF-D levels in tumor tissue may predict the effectiveness of bevacizumab treatment in combination with chemotherapy [23]. However, the value of VEGF-targeted therapy for patients with mCRC has not been established. Carbohydrate antigen 19-9 (CA19-9) [21] was first isolated from a colorectal cancer cell line, and it is a standard tumor marker for pancreatic cancer [8]. Although CA19-9
13
410
is overexpressed in 39 % mCRC patients [7], its levels alone are insufficient for monitoring colorectal cancer treatment according to American Society of Clinical Oncology guidelines [13]. CA19-9 acts as an adhesion factor for cells lining the blood vessels [21]. Furthermore, the association between CA19-9 levels and the efficacy of bevacizumab treatment remains unclear. The aim of this study was to determine the usefulness of pretreatment CA19-9 levels for predicting the survival and bevacizumab treatment efficacy in mCRC patients. We analyzed (1) the prognostic impact of baseline CA19-9 levels for mCRC patients receiving oxaliplatin-based first-line chemotherapy and (2) their ability to predict bevacizumab treatment efficacy.
Patients and methods
Cancer Chemother Pharmacol (2014) 73:409–416
serum CA19-9 level, KRAS/BRAF mutation status in tumor tissue, survival status, and date of last follow-up or death. To detect point mutations in KRAS codons 12 and 13, as well as the V600E mutation in the BRAF gene, we used the cycleave PCR technique [24]. The PCR reactions were performed using a cycleave PCR core kit (TAKARA, Co. Ltd., Ohtsu, Japan). Fluorescent signals were quantified using the Smart Cycler system (SC–100; Cepheid, Sunnyvale, CA, USA). Treatments The details of the treatment schedule are presented in Supplementary Data. Patients received a FOLFOX-4, modified FOLFOX-6, or CapeOX chemotherapy regimen [1, 5, 17]. Further, patients treated with bevacizumab [5, 17], cetuximab [14], and cediranib (AZD2171) [18] were included in this study. Bevacizumab has been approved for mCRC since April 2007 in Japan.
Patients Statistical analyses This study was a retrospective cohort study of mCRC patients who received oxaliplatin-based regimens (including FOLFOX-4, modified FOLFOX-6, and CapeOX) as first-line chemotherapy at Aichi Cancer Center Hospital. Patients were aged >20 years, had histologically proven inoperable colorectal cancer, and had serum CA19-9 levels available within 4 weeks of starting chemotherapy. Patients were excluded if they had double cancer or had previously undergone chemotherapy for advanced disease. CA19‑9 level measurement CA19-9 levels were measured at a certified laboratory in our institution using LUMIPULSE f® (Fujirebio, Inc.,) [15]. Reference range for normal CA19-9 levels was ≤37 U/ml [15]. Red cell phenotyping for Lewis antigen status was not performed in majority of the patients; therefore, patients, whose CA19-9 levels were persistently 70 years); gender (female vs. male); ECOG PS (0, 1 vs. 2, 3); primary tumor site (colon vs. rectum); number of metastatic sites (1 vs. ≥2); liver involvement (no vs. yes); prior colectomy (no vs. yes); histological subtypes (wel/mod vs. por/sig/muc); WBC count (50 ng/ml); and CA19-9 levels (>1, ≤37 vs. >37 IU/ml). A multivariate analysis using Cox proportional hazards model was performed to assess the relationship between clinicopathological variables and survival. To assess whether baseline CA19-9 is a prognostic marker independent of bevacizumab treatment and whether it predicts the patients’ response to bevacizumab, we separated patients into four groups on the basis of whether they underwent bevacizumab treatment and whether baseline CA19-9 levels were normal or elevated. We excluded the patients treated with cetuximab/cediranib from this analysis. We also assessed the relationship between baseline CEA and OS.
411
Cancer Chemother Pharmacol (2014) 73:409–416 Table 1 continued
Table 1 Patient characteristics Characteristics
N = 252
Characteristics
Age, years Median (range) Gender Female Male ECOG PS 0 1 2 3 Disease status Unresectable Recurrent Primary site Colon Rectum Number of metastatic sites 1 ≥2 Metastatic sites Liver Lung Lymph node Peritoneum Prior colectomy Yes No Histological subtypes Wel Mod Por Muc Sig Chemotherapy First-line FOLFOX/CapeOXa FOLFOX/CapeOXa + bevacizumab FOLFOX/CapeOXa + cetuximab FOLFOX/CapeOXa ± AZD2171 Second-line Third-line WBC (/μl) 37 IU/ml) Undetectableb (≤1 IU/ml)
143 (56.7) 109 (43.3) 159 (63.1) 93 (26.9) 79 (31.3) 150 (59.5) 23 (9.2)
KRAS/BRAF status 124 (49.2) 128 (50.8)
KRAS mutant G12X G13X
148 (58.7) 104 (41.3)
BRAF mutant V600E
116 (46.0) 136 (54.0) 118 (46.8) 146 (57.9) 153 (60.7) 193 (76.6) 210 (83.3) 42 (16.7) 55 (21.8) 166 (65.9) 22 (8.7) 5 (2.0) 4 (1.6)
133 (52.8) 101 (40.0) 8 (3.2) 10 (4.0) 190 (75.4) 99 (39.3) 237 (94.0) 15 (6.0) 128 (50.8) 124 (49.2)
KRAS wild type/BRAF wild type NA Outcome Median follow-up [range (months)] 30-day mortality Death Censored
34 (13.5) 10 (4.0) 7 (2.8) 76 (30.5) 124 (49.2) 23.3 (1.4–86.5) 0 (0) 197 (78.2) 55 (21.8)
ALP alkaline phosphatase, CA19-9 carbohydrate antigen 19-9, CEA carcinoembryonic antigen, ECOG Eastern Cooperative Oncology Group, LDH lactate dehydrogenase, Mod moderately differentiated adenocarcinoma, Muc mucinous carcinoma, NA not available, Por poorly differentiated adenocarcinoma, PS performance status, Sig signet ring cell carcinoma, WBC white blood cell count, Wel welldifferentiated adenocarcinoma a
FOLFOX: FOLFOX-4 or modified FOLFOX-6 (fluorouracil + leucovorin + oxaliplatin) b
Undetectable: Lewis antigen negative (Lea−b−)
Results were considered statistically significant when the two-sided P value was
