Chemotherapy-induced transient CEA and CA19-9 surges in patients with metastatic or recurrent gastric cancer

Abstract

Background. Rising serum tumor markers after chemotherapy are generally considered to indicate tumor progression. However, we have observed a transient increase in carcinoembryonic antigen (CEA) or carbohydrate antigen 19-9 (CA19-9) levels despite clinical benefits from chemotherapy in patients with metastatic or recurrent gastric cancer (MRGC). Therefore, this study was performed to determine the incidence of CEA and CA19-9 surges and their implications on the clinical outcome in MRGC patients. Material and methods. Fifty-one and 40 patients who had evaluable data for CEA or CA19-9 surges, respectively, were included. Both CEA and CA 19-9 surges were defined as a > 20% increase in these tumor markers from the baseline that was followed by a > 20% drop in subsequent levels compared to the baseline. Results. Of 51 evaluable patients for CEA surges, nine patients (18%) had documented CEA surges. The median time to CEA peak and the duration of the CEA surge were 2.8 (1.7∼7.0) and 9.1 weeks (7.6∼21.0), respectively. Of 40 evaluable patients for CA19-9 surges, seven patients (18%) had CA19-9 surges. The median time to peak and the duration of the CA19-9 surge were 2.3 (1.9∼6.0) and 7.1 weeks (4.3∼16.1), respectively. All patients with CEA or CA19-9 surges had radiographic evidence of benefits from chemotherapy. Conclusion. CEA or CA19-9 surges can be observed in MRGC patients receiving chemotherapy. All patients with these surge phenomena had clinical benefits from chemotherapy. An initial rise in CEA or CA19-9 levels after initiation of chemotherapy should not be used as an indicator of progressive disease.

Although the incidence of gastric cancer has decreased significantly in Western countries, it is still among the most common malignancies in South America, in many former Eastern European countries, and across Asia. The incidence of gastric cancer is the second leading cause of cancer mortality in the world. In Korea, according to statistics reported in 2005, gastric cancer is the most prevalent cancer in males [1]. Randomized trials have found that cytotoxic chemotherapy is associated with an improvement in overall survival (OS) and quality of life as compared with best supportive care [2]. Recently, 5-fluorouracil (5-FU) and platinum-based combination chemotherapy has been considered as the standard regimen in the first-line setting for metastatic or recurrent gastric cancer (MRGC) [3–6], and other drugs including taxane and irinotecan are also commonly used [7–9].

Baseline levels of carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (CA19-9) are commonly observed to increase before the initiation of chemotherapy in MRGC. CEA is a β-1-glycoprotein normally found in the gastrointestinal tract of embryos, and in smaller concentrations, in normal adult tissues. CEA functions as an intercellular adhesion molecule promoting the aggregation of human gastrointestinal cancer cells [10]. Carbohydrate antigen 19-9 (CA19-9) is a predominantly carbohydrate antigen which was defined from the culture medium of a colorectal cancer (CRC) cell line. It is a high molecular weight glycolipid derived from monoclonal antibodies isolated from mice which have been immunized with a human colon cell line [11]. During chemotherapy, the response to treatment is estimated by radiographic evaluation, but if radiologic assessment is difficult or obscure, the change in tumor marker levels may be used as adjuvant means of treatment monitoring [12]. Also, many clinicians try to predict the effect of chemotherapy by obtaining serial levels of tumor markers during chemotherapy before follow-up with imaging, such as computed tomography (CT) scan. In general, a rising tumor marker level means tumor progression in patients who are receiving chemotherapy. However, in the case of CEA, transient elevations followed by a decrease below the initial baseline level after initiation of chemotherapy have been observed in metastatic colorectal and breast cancer patients who benefited from palliative chemotherapy [13–18]. This phenomenon is called the ‘CEA surge,’ and these initial increases in tumor marker levels during chemotherapy can confuse clinicians with resultant errors in the determination of benefits from the chemotherapy. In the case of CA19-9, it has been generally considered as an important marker for response to treatment, as well as for prognosis in pancreatic cancer [19]. However, although guidelines of the American Society of Clinical Oncology (ASCO) allow the use of CA19-9 during active treatment period for metastatic pancreatic cancer [20], data supporting routine CA19-9 measurements during chemotherapy for monitoring the clinical courses of patients with metastatic pancreatic cancer are not solid. A recent study showed that CA19-9 response during chemotherapy was not a valid surrogate marker for survival in pancreatic cancer patients [21]. Therefore, the routine use of CA19-9 in various solid tumors, including pancreatic cancer, is still controversial. Additionally, to our knowledge, a ‘CA19-9 surge’ phenomenon has not been reported.

In MRGC patients, tumor marker surge phenomena have not been described. Moreover, ASCO guidelines for the use of tumor markers in gastrointestinal cancer have not discussed the use of tumor markers in gastric cancer [20]. Therefore, this study was performed to determine the incidence of CEA and CA19-9 surges and their implications for clinical practice in MRGC patients.

Patients and methods

Study population and tumor response evaluation

Using databases, MRGC patients who had received palliative chemotherapy at Seoul National University Bundang Hospital were retrospectively analyzed. From January 2005 to October 2007, a total of 177 MRGC patients received chemotherapy. The number of patients with increased baseline CEA and CA19-9 levels before the initiation of chemotherapy were 70 and 63, respectively. Fifty-one and 40 patients had serial tumor marker follow-up data for CEA and CA19-9 during chemotherapy, respectively, and measurable lesions for the response evaluation to chemotherapy and thus were included in this study. In these enrolled patients, CEA or CA19-9 level tests were carried out before each cycle of chemotherapy. Computed tomography (CT) scans were performed to document the disease extent and to evaluate the response to treatment every 6∼9 weeks according to the administered chemotherapy regimens. Response was assessed using the Response Evaluation Criteria in Solid Tumors (RECIST) [22]. All patients gave written informed consent before initiating treatment and this study was approved by the Institutional Review Board.

Serum assays for CEA and CA19-9

Serum CEA and CA 19-9 were measured by an electrochemiluminescence immunoassay which used a cobas e immunoassay analyzer (Roche Diagnostics, Mannheim, Germany). The measurement range was 0.2 – 1 000 ng/ml, and the specimen was diluted by an automated dilution protocol (1:50) if exceeding the upper limit value. The recommended cut-off levels of serum CEA and CA 19-9 were 5 ng/ml and 37 U/ml, respectively.

Definition of CEA and CA 19-9 surge

The definition of CEA progression varies from a>10% rise from baseline value to a 20∼50% increase [16], [17]. Two recent studies, which demonstrated a ‘CEA surge’ phenomenon in metastatic CRC, used the criterion of ‘a>20% increase from the baseline CEA value followed by a>20% drop in subsequent CEA levels compared to the baseline value’ [17], [18]. We also used the same criterion for CEA surge in this study. However, there have been no reports about ‘CA19-9 surge’ phenomenon. Therefore, we arbitrarily applied the same criterion to CA19-9 surge as used with CEA. The measurement of baseline CEA and CA19-9 levels was taken<2 weeks before the start of chemotherapy.

Results

Patient characteristics

Among 177 patients reviewed, only patients with measurable lesions, increased baseline tumor marker levels, and serial tumor marker follow-up data were included in this study. Median age of all patients was 61 years (range: 21-88). Fifty one and 40 patients were evaluable for CEA and CA19-9 changes during chemotherapy, respectively; the characteristics of the patients are shown in Table I. The incidences of CEA and CA19-9 surges were 18% (N=9 [95% confidence interval (CI): 7-28%]) and 18% (N=7 [95% CI: 6-29%]), respectively.

Table I.  Characteristics of eligible patients and patients with CEA or CA 19-9 surges.

Variable	Patients with increased baseline CEA levels (No.)	Patients with a CEA surge (No. [%])	Patients with increased baseline CA 19-9 levels (No.)	Patients with a CA19-9 surge (No. [%])
Overall	51	9 (18)	40	7 (18)
Age
<60 years	22	6 (27)	19	1 (5)
≥60 years	29	3 (10)	21	6 (29)
Gender
Male	37	6 (16)	26	6 (23)
Female	14	3 (21)	14	1 (7)
Metastasis
Liver	14	3 (21)	13	3 (23)
Peritoneum	24	3 (13)	19	1 (5)
Bone	15	3 (20)	8	1 (13)
Lymph node	33	3 (9)	25	6 (24)
Previous gastrectomy
Yes	17	2 (12)	17	2 (12)
No	34	7 (21)	23	5 (23)

CEA surge

Various chemotherapeutic regimens were administered in 9 patients with CEA surges (Table IIA). Of these patients who showed transient CEA elevations, one patient had CEA surge twice during the third- (irinotecan, 5-FU, and leucovorin [FOLFIRI]) and forth-line (paclitaxel and carboplatin) treatment periods. From the time of MRGC diagnosis, this patient has consistently shown good responses to consecutive palliative chemotherapeutic regimens and has survived for>3 years. All nine patients (10 surge events) with CEA surges benefited from the administered chemotherapy; all radiologic response evaluations in these patients showed partial remission (PR) or stable disease (SD) (Table IIA). The distribution of CEA values in patients with a CEA surge is also shown in Table IIA. The median baseline CEA levels before the initiation of chemotherapy and peak CEA levels during chemotherapy were 29.2 ng/ml (range, 5.4-1900 ng/ml) and 55.7 ng/ml (range, 8.4-3100 ng/ml), respectively. The median time to CEA peak from the start of chemotherapy was 2.8 weeks (range, 1.7-7.0 weeks). The CEA surge lasted a median 9.1 weeks (range, 7.6-21.0 weeks).

Table II.  Time to peak, duration of tumor marker surges (the period between the initiation of chemotherapy and a>20% drop below the baseline level), and clinical outcome in patients with tumor marker surges; (A) CEA surge and (B) CA19-9 surge [Abbreviations: PR, partial disease; SD, stable disease; FOLFIRI, irinotecan+5-FU+leucovorin; Pac, paclitaxel; Carbo, carboplatin; Cis, cisplatin; FOLFOX, oxaliplatin+5-FU+leucovorin; Doc, docetaxel; Ctx, cetuximab; 1^st, first-line chemotherapy; 2^nd, second-line chemotherapy; 3^rd, third-line chemotherapy; and 4^th, fourth-line chemotherapy].

Case	Response	Chemotherapy regimen	Baseline CEA (ng/ml)	Peak CEA (ng/ml)	Nadir CEA (ng/ml)	Time-to-peak (weeks)	Duration of surge (weeks)
1	PR	FOLFIRI (3^rd)	25.0	49.6	19.4	5.4	21
2	PR	Pac/Carbo (4^th)	46.5	73.1	15.3	2.9	15.9
3	PR	TS-1/Cis (1^st)	1900	3100	214	2.6	9.6
4	PR	FOLFOX (1^st)	16.8	20.7	8.3	2.0	8.1
5	PR	TS-1/Cis (1^st)	5.4	8.4	3.4	2.7	7.7
6	PR	FOLFOX (1^st)	165	466	51.6	1.7	8.1
7	SD	FOLFIRI (4^th)	434	522	271	5.0	7.6
8	SD	FOLFIRI (3^rd)	15.2	33.6	10.5	1.9	8.6
9	SD	FOLFIRI (2^nd)	22.9	32.7	11.7	7.0	10.9
10	SD	Doc/Cis (2^nd)	33.3	61.7	24.9	3.9	17.4
(A)
Case	Response	Chemotherapy regimen	Baseline CA 19-9 (U/ml)	Peak CA19-9 (U/ml)	Nadir CA19-9 (U/ml)	Time-to-peak (weeks)	Duration of surge (weeks)
1	CR	FOLFOX (1^st)	81	105	40	1.9.	5.7
2	PR	TS-1/Doc (1^st)	57	73	41	3.0	13.3
3	PR	TS-1/Cis (1^st)	514	930	220	2.3	7.6
4	PR	TS-1/Cis (1^st)	13300	20000	3900	2.6	5.9
5	PR	Doc (3^rd)	41	83	33	6.0	16.1
6	PR	FOLFOX (1^st)	276	370	85	2.1	7.1
7	PR	Ctx/FOLFOX (1^st)	1610	2300	850	2.0	4.3

(B)

CA 19-9 surge

The description of patients with a CA19-9 surge and the distribution of CA19-9 values are shown in Table IIB. Various chemotherapeutic regimens showed CA19-9 surge phenomena. All seven patients with a CA19-9 surge had clinical benefits from chemotherapy; all patients achieved complete remission (CR) or PR. The median baseline and peak CA 19-9 levels were 276 U/ml (range, 41-13300 U/ml) and 370 U/ml (range, 73-20000 U/ml), respectively. The median time to CA19-9 peak was 2.3 weeks (range, 1.9∼6.0 weeks) from the start of chemotherapy and the surge duration lasted a median 7.1 weeks (range, 4.3-16.1 weeks).

Discussion

Transient tumor marker elevations in patients who benefited from chemotherapy have been observed in a few types of solid tumors. A surge of human chorionic gonadotropin (HCG) and α-fetoprotein (AFP) has also been described in non-seminomatous germ cell tumor (NSGCT). During the initial weeks of treatment, a transient increase in either or both markers is seen in 40 – 70% of NSGCT patients [23], [24]. In breast cancer, serial CEA follow-up during chemotherapy showed initial rising and then declining CEA levels in 11 – 48% of patients responding to chemotherapy [13–16]. CEA surges are also commonly observed in patients with CRC [17], [18]. Sørbye et al [17]. serially measured CEA after 5-FU, oxaliplatin, and folinic acid (FOLFOX) chemotherapy. Of 27 patients with metastatic CRC, four patients (15%) met the criteria of a CEA surge, and the time to reach the peak CEA level varied between 13 and 56 days. Ailawadhi et al [18]. also observed a CEA surge in ten (11%) of 89 CRC patients after 5-FU-based chemotherapy. The median time to CEA peak was 4 weeks (2-10 weeks), and the CEA surges lasted 4 – 16 weeks.

However, to our knowledge, transient CEA surge phenomena in MRGC patients responding to chemotherapy have not been previously described. In the case of CA19-9, a surge phenomenon has never been published in solid tumors. In the current study, we checked CEA and CA 19-9 levels serially in MRGC patients being treated with systemic chemotherapy and demonstrated a CEA surge in 9 patients (18%) and CA 19-9 surge phenomena in seven patients (18%). The median time to peak levels of CEA and CA19-9 was 2.8 (range, 1.7-7.0) and 2.3 weeks (range, 1.9-6.0 weeks), respectively. The median duration of CEA and CA19-9 surges was 9.1 (range, 7.6-21.0) and 7.1 weeks (range, 4.3-16.1 weeks), respectively.

Although serum tumor markers have not been established as the parameter for monitoring response to chemotherapy in MRGC, some studies have evaluated CEA and CA 19-9 as surrogate markers of response to systemic chemotherapy and found a significant correlation between the decrease or increase of tumor markers with radiologic response [12], [25]. Although the RECIST criterion is the current standard for response evaluation to chemotherapy [22], a sizable number of cancer patients do not have any measurable lesions and thus the RECIST criterion can not be applied. In MRGC cases, patients only with unmeasurable lesions, such as peritoneal, bone, or lymphangitic lung metastases, are common. For example, in large scale phase III studies, 25 – 35% of MRGC patients had no target lesion and thus response measurement during chemotherapy in these patients could not be performed according to the RECIST criteria [4], [26]. Our observations on CEA or CA19-9 surge phenomena and the time courses of surges will provide clinicians with auxiliary tools in decision-making on maintaining or discontinuing the current chemotherapeutic regimen in MRGC patients.

The mechanism of surge phenomenon is unknown, but several hypotheses can be analogized. First, chemotherapeutic agents can induce an increase in tumor marker expression by cancer cells, resulting in an initial surge with subsequent decline in association with enhanced tumor cell kill. This would be supported by preclinical data showing up-regulated expression of CEA in tumor cell lines in response to 5-FU or platinum exposure [18], [27], [28]. Second, tumor cell lysis may explain surge phenomena. The initial tumor marker surge may be due to rapid tumor cell lysis in treatment responders [17].

In conclusion, an initial rise in CEA or CA19-9 levels during chemotherapy in MRGC may not indicate tumor progression, but may represent a transient tumor marker surge phenomenon in patients responding to treatment. When using tumor markers as part of the monitoring of response to chemotherapy, treatment decisions based on a rise in tumor marker levels should be deferred unless CA19-9 or CEA levels continuously increase>6 or 7 weeks.

Conflict of interest

We declare no conflicts of interest.