Original Articles
Copyright ©The Author(s) 1998. Published by Baishideng Publishing Group Inc. All rights reserved.
World J Gastroenterol. Apr 15, 1998; 4(2): 147-149
Published online Apr 15, 1998. doi: 10.3748/wjg.v4.i2.147
A clinical evaluation of serological diagnosis for pancreatic cancer
Xiao-Yan Zhao, Shi-Yuan Yu, Shi-Ping Da, Li Bai, Xiao-Zhong Guo, Xiao-Jing Dai, Yuan-Ming Wang
Xiao-Yan Zhao, Shi-Yuan Yu, Shi-Ping Da, Li Bai, Xiao-Zhong Guo, Xiao-Jing Dai, Yuan-Ming Wang, Department of Gastroenterology, Xinqiao Hospital, Third Military Medical University, Chongqing 430037, China
Xiao-Yan Zhao, Ph.D., male, born on 1962-03-29 in Wuhan City, Hubei Province, graduated from the Third Military Medical University as a postgraduate in 1995, now associated professor and assistant director of the Department of Gastroenterology, having 12 papers published.
Author contributions: All authors contributed equally to the work.
Correspondence to: Dr. Xiao-Yan Zhao, Department of Gastroenterology, Xinqiao Hospital, Third Military Medical University, Chongqing 430037, China
Telephone: +86·23·68755604
Received: December 5, 1997
Revised: December 30, 1997
Accepted: January 12, 1998
Published online: April 15, 1998

Abstract

AIM: To assess the diagnostic values of tumor markers for pancreatic cancer.

METHODS: Pancreatic cancer-associated antigen from colonic mucosa (PCAAc), pancreas-specific antigen (PaA), pancreatic oncofetal antigen (POA) and minimolecular pancreatic antigen (mPOA) were detected by double antibodies Sandwich ELISA; CA19-9, elastase 1 (E1), human pancreatic elastase 1 (HPE1) and carcinoembryonic antigen (CEA) by radioimmunoassay (RIA); general activities of ribo-nuclease (RNase) and its isoenzymes (RNase I and RNase II) by biochemistry and PAEG; glycylproline dipeptidyl aminopeptidase (GPDA) by biochemistry and α1-antitrypsin (α 1AT) by rocket immunoelectrophoresis (rocket-IE).

RESULTS: The detection of serum POA, mPOA, PaA, PCAAc, CA19-9, RNase and RNase I was able to differentiate pancreatic cancer from the benign disorders and non-pancreatic malignancies with a sensitivity from 66.75% to 80.0% and a specificity from 88.5% to 96.69%. POA, mPOA, PCAAc, HPE1, E1 and GPDA were related to the pancreatic cancer at the head which demonstrated higher sensitivity from 63.64% to 85.71%. The detection of serum HPE1 was especially helpful for the diagnosis of pancreatic cancer with smaller diameters. The determination of 3 or 4 kinds of tumor markers simultaneously would increase the detection rate of pancreatic cancer, which will be an important procedure for the diagnosis of this malignancy.ª¤

CONCLUSION: A single test of tumor markers is helpful to detect pancreatic cancerclinically, but the determination of 3 or 4 kinds of tumor markers simultaneously would significantly increase the detection rate of pancreatic cancer, which will be an important procedure for the diagnosis of this malignancy.

Key Words: pancreatic neoplasms/diagnosis, tumor markers, biological, antigens, neoplasm/analysis, CA-19-9 antigen/analysis, pancreatopeptidase/analysis, carcinoembryonic antigen/analysis, alpha 1-antitrypsin/analysis, enzyme-linked immunosorbent assay, radioimmunoassay



INTRODUCTION

The incidence of pancreatic cancer is increasing yearly, but its early diagnosis remains very difficult, for the symptoms are faint, vague and nonspecific. Therefore, it is important to seek a tumor marker with satisfactory specificity and/or sensitivity for the diagnosis of the cancer clinically. This paper reports our ten years’ research on tumor markers to evaluate the clinical application of the serum detection of tumor markers for the diagnosis of pancreatic cancer.

MATERIALS AND METHODS
Samples

We detected 13 tumor markers in samples from 763 cases including 143 pancreatic cancer, 125 liver or peri-ampullar carcinoma and 120 gastrointestinal carcinoma, which were identified by surgical operation and/or pathological examination. One hundred and five cases of acute and chronic pancreatitis were observed. The acute cases were identified by clinical approach and chronic ones by X-ray photograph, surgical operation and pathological examination. Blood samples from 270 healthy fasting donors were examined likewise and served as control. Fasting blood samples were obtained through venous puncture in the morning and the serum was preserved at -20 °C for assay.

Methods

Pancreatic cancer-associated antigen from colon mucosa (PCAAc), pancreas-specific antigen (PaA), pancreatic oncofetal antigen (POA) and minimolecular pancreatic antigen (mPOA) was detected by double antibodies Sandwich ELISA[1-4]; CA19-9, elastase 1 (E1), human pancreatic elastase 1 (HPE1) and carcinoembryonic antigen (CEA) by radioimmunoassay (RIA); general activities of ribonuclease (RNase) and its isoenzymes (RNase I and RNase II) by biochemistry and PAEG[5]; glycylproline dipeptidyl aminopeptidase (GPDA) by biochemistry according to Kojima’s report[6] and α1-antitrypsin (α1AT) by rocket immunoelectrophoresis (rocket-IE).

Statistical analysis

The data of distribution percentage, t test, FLSD test, stepwise regression and discriminatory analysis of the tumor markers were processed on a microcomputer with a medical program system, SPMR, designed by the department of mathematics of our college.

RESULTS
Single assay of the tumor markers

Circulating levels of the tumor markers in patients with pancreatic cancer as well as those with benign pancreatic diseases, non-pancreatic malignancies and normal control subjects are shown in Table 1 and Table 2. All the tumor markers except for RNase 2 and CEA were elevated in the sera of patients with pancreatic cancer and exhibited a sensitivity ranging from 45% to 80.8% and a specificity from 66.67% to 97.14%.

Table 1 Sensitivity and specificity of a single test of tumor markers for pancreatic cancer.
GroupsSensitivity (%)Specificity (%)
POA67.8088.50
mPOA74.3695.6
PCAAc68.7096.49
PaA78.7989.40
HPE166.6783.25
E162.5066.67
CA19-970.9790.63
GPDA42.1268.42
RNase80.8080.00
RNase I45.0094.28
RNase II25.0097.14
α1AT50.0074.28
CEA25.0085.71
Table 2 Results of single test of tumor markers.
GroupsNormal controlPancreatic cancerNon-pancreaticmali gnancyPancreatitis
POA (U/mL)84.80 ± 13.96 (1/40)252.5 ± 164.33 (19/28)136.37 ± 93.63 (9/62)101.98 ± 57.88 (1/15)
mPOA (mg/L)4.88 ± 1.96 (2/40)18.45 ± 10.57 (29/39)6.24 ± 3.82 (5/62)4.69 ± 1.52 (0/17)
PCAAc (mg/L)12.59 ± 6.34 (1/40)57.25 ± 82.93 (23/33)12.64 ± 5.14 (3/49)13.53 ± 9.91 (0/14)
PaA (μg/L)13.05 ± 3.84 (3/64)61.54 ± 45.59 (26/33)17.62 ± 10.31 (5/44)21.35 ± 11.83 (3/20)
HPE1 (μg/L)23.80 ± 3.40 (1/82)83.30 ± 73.40 (62/93)48.93 ± 54.63 (26/63)124.90 ± 168.30 (25/35)
E1 (ng/dL)276.60 ± 85.36 (0/10)1116.00 ± 1281.00 (10/16)1321.00 ± 2842.00 (9/13)
CA19-9 (U/L)7.90 ± 9.74 (2/81)165.40 ± 128.03 (19/31)139.88 ± 121.72 (6/15)
GPDA (U/L)54.84 ± 10.80 (0/60)101.02 ± 75.84 (14/33)64.89 ± 72.64 (27/69)50.94 ± 15.50 (2/25)
RNase (U/mL)58.19 ± 11.09 (1/30)92.22 ± 19.66 (16/20)65.88 ± 12.32 (5/20)67.95 ± 9.77 (2/15)
RNase I (U/mL)9.53 ± 13.43 (1/30)39.09 ± 39.31 (9/20)9.23 ± 8.98 (1/20)11.87 ± 11.30 (1/15)
RNase II (U/mL)50.75 ± 13.36 (0/30)52.78 ± 29.68 (5/20)56.89 ± 16.82 (0/20)60.17 ± 19.55 (1/15)
α1AT (mg/L)30.51 ± 6.35 (1/30)49.86 ± 18.47 (10/20)38.65 ± 16.30 (6/20)39.23 ± 10.02 (3/15)
CEA (mg/L)11.15 ± 2.95 (1/20)16.82 ± 10.26 (5/20)15.73 ± 11.98 (5/20)10.22 ± 3.42 (0/15)
Tumor markers related to location, size and differentiation of pancreatic cancer

As shown in Table 3, the detection of POA, mPOA, PCAAc, E1, HPE1 and GPDA demonstrated that the positive percentage ranged from 63.64% to 85.73% in the patients with pancreatic cancer at the head, which was significantly higher than that at the body or tail and entire. The detection of PaA and CA19-9 could be better for the cancer at tail and/or body. It is worth-mentioning that higher positivity of the detection with HPE1 was found in the patients with the tumor smaller than 5.0 cm. The elevation of serum mPOA, PCAAc, PaA, HPE1, E1 and GPDA was closely related to well-differentiated cancer.

Table 3 The relationship between the tests of tumor markers and the location, size or differentiation of pancreatic cancer.
GroupsLocation
Size
Differentiation
HeadBody or tailEntire< 5.0 cm5 cm-9.0 cm< 9.0 cmWell & moderatePoor & un-differentiated
POA9/11 (81.82)3/6 (50.00)1/3 (33.33)4/8 (50.00)8/10 (80.00)1/2 (50.00)
mPOA22/27 (81.48)5/9 (55.56)2/3 (66.67)5/5 (100.00)14/19 (73.68)10/15 (66.67)10/10 (100.00)9/13 (69.23)
PCAAc17/22 (77.27)4/9 (44.44)1/2 (50.55)3/4 (75.00)9/13 (69.23)3/6 (50.00)5/5 (100.00)2/4 (50.00)
PaA16/22 (72.73)10/11 (90.91)2/2 (100.00)4/4 (100.00)12/15 (80.00)6/10 (60.00)8/8 (100.00)4/5 (80.00)
E114/17 (82.35)2/6 (33.33)2/3 (66.67)6/8 (75.00)4/10 (40.00)1/2 (50.00)3/6 (50.00)2/4 (50.00)
HPE135/41 (85.37)20/35 (57.14)7/17 (41.18)40/51 (78.43)15/25 (60.00)7/17 (41.18)43/60 (71.67)19/33 (57.58)
GPDA14/22 (63.64)0/9 (0.00)0/2 (0.00)8/17 (47.06)3/7 (42.86)3/7 (42.86)4/7 (57.14)1/4 (25.00)
RNase31/41 (75.61)24/35 (68.23)15/17 (88.23)20/29 (68.96)19/25 (76.00)15/17 (88.24)46/ 60 (76.67)24/33 (72.72)
CA19-910/17 (58.82)6/8 (75.00)6/6 (100.00)7/15 (46.67)9/12 (75.00)4/4 (100.00)8/9 (88.89)8/8 (100.00)
Simultaneous detection of tumor markers

Although single assay of tumor markers distinguishes pancreatic cancer from the other groups, the conjoint detection could improve the diagnostic positivity and specificity for the cancer clinically. As shown in Table 4, four tumor markers, including PCAAc, mPOA, PaA and GPDA were selected for this purpose, and the results showed a positivity of 40.6%-80.8% for the diagnosis of pancreatic cancer by single test, 75%-96.77% by conjoint detection with any two markers and 100% by any three markers.

Table 4 Results of simultaneous detection of four tumor markers.
GroupsPancreatic cancer (n = 30)Non-pancreatic malignancy (n = 43)Benign diseases (n = 21)
All the tests negative0 (0.00)17 (39.54)18 (85.71)
Single test positive4 (12.90)21 (48.83)3 (14.29)
Double tests positive6 (19.36)5 (11.63)0 (0.00)
Three tests positive16 (51.61)0 (0.00)0 (0.00)
All the tests positive5 (16.13)0 (0.00)0 (0.00)
DISCUSSION
Single assay significance of tumor markers for pancreatic cancer diagnosis

A single assay of tumor markers was helpful in distinguishing patients with pancreatic cancer from normal controls and those with non-pancreatic malignancy. By stepwise discriminatory analysis, seven out of all 13 tumor markers including CA19-9, POA, mPOA, PaA, PCAAc, RNase and RNase I were associated well with the diagnosis of pancreatic cancer with the sensitivities ranging from 45% to 80.8% and the specificity from 80% to 96.49%. Although the detection of HPE1 and E1 had the sensitivities of 66.67% and 62.5% to the disease, these were regarded as helpful diagnostic indices for the poor differentiation value. A single test of GPDA, α1AT, CEA and RNase II was not significant to the diagnosis of the cancer for their unsatisfied sensiti-vity and/or specificity.

Although the mechanism of the tumor markers expressed in blood was not clear, the problem of false positive rate was noticed in clinic; especially to some carcinoma adjacent to pancreatic cancer such as primary hepatic carcinoma, cholangioampulla carcinoma, etc.[7,8]. In our study, the false positive rates ranging from 8.33% to 21.74% was found in patients with primary hepatic carcinoma or cholan-gioampulla carcinoma with a single test of POA, mPOA, PCAAc, PaA and RNase I, and 36.92% to 87.09% with CA19-9, HPE1, E1 and GPDA. These suggest that a single test of tumor markers has limitations for making a definite diagnosis of pancreatic cancer unless it is combined with other detection, especially the image analysis.

Correlation of tumor markers to location, size or differentiation of cancer

The higher positive rates in detecting POA, PCAAc, HPE1, E1 or GPDA were found in patients with pancreatic head cancer, with the sensitivities of 63.64%-85.71%. It is interesting that GPDA were only elevated in the sera of patients with the cancer at pancreatic head but not at body or/and tail. The mechanism of elevated GPDA might be related to the obstruction of bile duct. Unfortunately, none of the 13 tumor markers were significantly related to the cancer at body or/and tail.

According to our studies, the sensitivities of CA19-9, RNase, GPDA and POA detection were gradually increased with the development of the mass, which were 46.67%-50% in the patients with the tumor size less than 5.0 cm. Although elevated serum mPOA and PCAAc were associated with the patients with the tumor size less than 5.0 cm, the methods per se for correct diagnosis of pancreatic cancer in the early stage still need further studies. It is worth-mentioning that the detection of serum HPE1 was significantly related to the tumors less than 5.0 cm. In this study, significantly higher serum HPE1 was observed in 22 of 51 cases with the tumor size less than 3.0 cm, among them the concentration was 118.1 µg/L ± 102.7 µg/L and the positive rate was 78.46%. Furthermore, 4 of 6 cases with the tumor size less than 2.0 cm also had elevated serum HPE1. This suggests that serum HPE1 might be a reference tumor marker helpful for the diagnosis of patients in the early stage.

Correlation of tumor markers with differentiation of cancer

The serum variation of some tumor markers such as mPOA, PCAAc, PaA, HPE1 and GPDA were related to the differentiation degree of the tumors. Our results showed that positive rate and elevation range in highly and moderately differentiated pancreatic cancer were higher than low and non-differentiated one, which suggested that these tumor markers can be regarded as differentiating antigen for assessment of the transformation of the disease.

Significance of combined assay of tumor markers for pancreatic cancer

Combined detection of tumor markers would increase and improve the detection rate and the diagnostic accuracy of the malignancy. According to our data, the sensitivity to the diagnosis of the disease was 25%-80.0% with single test of the 13 tumor markers, 75%-96.77% with two tests, and 100% with over 3 tests. The combined tests with simultaneous positive results of 2 out of the 13 tumor markers could be used to rule out pancreatic benign disease. Non-pancreatic malignancies never showed positive results of over three tumor markers simultaneously. Our results suggest that determination of 3 or 4 tumor markers conjointly would increase the detection rate of pancreatic cancer, especially for screening among large populations. It might become an important procedure for the diagnosis of this malignancy.

Footnotes

Project supported by “8th 5-year” Major Scientific Research Foundation of Chinese PLA.

References
1.  Guo XZ, Yu SY, Guo P, Dai XJ, Liu WW. The value of measurement of serum pancreatic oncofetal antigen in the diagnosis of pancreatic carcinoma. Chin J Dig. 1989;9:88-91.  [PubMed]  [DOI]
2.  Zhao XY, Yu SY, Guo P, Bai L. Preliminary study of the diagnosis of pancreatic cancer with a pancreatic associated antigen. J Med Coll PLA. 1995;10:162-165.  [PubMed]  [DOI]
3.  Zhao XY, Yu SY, Bai L. [Diagnosis of pancreatic cancer with simultaneous detection of a pancreatic-cancer-associated antigen and pancreas-specific antigen]. Zhonghua Neike Zazhi. 1990;29:612-615, 612-615.  [PubMed]  [DOI]
4.  Dai XJ, Yu SY, Guo P, Guo XZ. A preliminary study on serum pancreatic elastase I for the diagnosis of pancreatic cancer. Chin J Dig. 1989;9:299-300.  [PubMed]  [DOI]
5.  Wang YM, Yu SY, Guo P. A preliminary study on serum ribonuclease and isoen-zyme for the diagnosis of pancreatic cancer. Chin J Dig. 1988;8(3): 134-137.  [PubMed]  [DOI]
6.  Kojima J, Kanatani M, Nakamura N, Kashiwagi T, Tohjoh F, Akiyama M. Serum and liver glycylproline dipeptidyl aminopeptidase activity in rats with experimental hepatic cancer. Clin Chim Acta. 1980;107:105-110.  [PubMed]  [DOI]
7.  Tian F, Appert HE, Myles J, Howard JM. Prognostic value of serum CA 19-9 levels in pancreatic adenocarcinoma. Ann Surg. 1992;215:350-355.  [PubMed]  [DOI]
8.  Saito S, Taguchi K, Nishimuri N, Watanabe A, Ogoshi K, Niva M et al. Clini-cal usefulness of computer associated diagnosis using combination assay of tumor markers for pancreatic carcinoma. Cancer. 1993;72:381-388.  [PubMed]  [DOI]