Scandinavian Journal of Gastroenterology
ISSN: 0036-5521 (Print) 1502-7708 (Online) Journal homepage: http://www.tandfonline.com/loi/igas20
Serum Carcinoembryonic Antigen and DNA Ploidy in Colorectal Carcinoma M. Kouri, S. Pyrhönen, J.-P. Mecklin, H. Järvinen, A. Laasonen, K. Franssila, P. Kuusela & S. Nordling To cite this article: M. Kouri, S. Pyrhönen, J.-P. Mecklin, H. Järvinen, A. Laasonen, K. Franssila, P. Kuusela & S. Nordling (1991) Serum Carcinoembryonic Antigen and DNA Ploidy in Colorectal Carcinoma, Scandinavian Journal of Gastroenterology, 26:8, 812-818, DOI: 10.3109/00365529109037017 To link to this article: http://dx.doi.org/10.3109/00365529109037017
Published online: 08 Jul 2009.
Submit your article to this journal
Article views: 2
View related articles
Citing articles: 2 View citing articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=igas20 Download by: [New York University]
Date: 15 March 2016, At: 04:40
Serum Carcinoembryonic Antigen and DNA Ploidy in Colorectal Carcinoma A Prospective Study
Downloaded by [New York University] at 04:40 15 March 2016
M. KOURI, S. PYRHONEN, J.-P. MECKLIN, H. JARVINEN, A. LAASONEN, K. FRANSSILA, P. KUUSELA & S. NORDLING Dept. of Radiotherapy and Oncology and Second Dept. of Surgery, University Central Hospital; Pathology Laboratory of the Dept. of Radiotherapy and Oncology and Dept. of Bacteriology and Immunology, Helsinki University; and Dept. of Pathology, Helsinki University; Helsinki, Finland Kouri M, Pyrhonen S, Mecklin J-P, Jarvinen H, Laasonen A, Franssila K, Kuusela P, Nordling S. Serum carcinoembryonic antigen and DNA ploidy in colorectal carcinoma. A prospective study. Scand J Gastroenterol 1991, 26, 812-818 We have analysed the relationship between carcinoembryonic antigen (CEA) and DNA ploidy prospectively in 130 colorectal carcinoma patients. CEA was elevated preoperatively significantly more often in patients with DNA-aneuploid tumours than in DNA-diploid or DNA-tetraploid tumours-that is, in 48% (36 of 7 5 ) of patients with aneuploid tumours, in 34% (14 of 41) of patients with diploid tumours, but only in 14% (2 of 14) of patients with tetraploid tumours ( p < 0.05). Aneuploid tumours had an elevated CEA level in 38% of stage A-B disease and in 61% of stage C-D disease. The elevated CEA values (25.0 &I) correlated with tumour stage in patients with aneuploid tumours but not in patients with diploid tumours. Whereas CEA is a suitable marker for aneuploid carcinomas, other more sensitive tumour markers should be sought for diploid and also for tetraploid tumours. If such markers are found, flow cytometry could provide the most important information in selecting individual follow-up programmes for colorectal cancer patients. Key words: Carcinoembryonic antigen; colorectal carcinoma; DNA-ploidies; flow cytometry Mauri Kouri, M . D . , Dept. of Radiotherapy and Oncology, Helsinki University Central Hospital, SF-00290 Helsinki, Finland
Carcinoembryonic antigen (CEA), first described by Gold & Freedman (l),is clinically the most important marker of colorectal cancer. The serum CEA assay does not possess the sensitivity or the specificity required to distinguish between localized malignant tumours and benign disorders (2). In a recent report 26% of the patients with resectable primary colorectal carcinoma and 72% of the patients with unresectable or metastatic disease were CEA-positive when a cut-off value of 5.0 pg/1 was used (3). CEA is very useful in the postoperative monitoring of radically operated patients. The CEA assay can detect about 60% to 90% of recurrences in an asymptomatic stage regardless of the CEA value at the time of
operation of the primary tumour, and up to about half of such patients can be reoperated on with curative intent (4-7). On the other hand, all patients with recurrence do not show a CEA increase, for reasons that need further research. There is an increasing number of studies relating prognosis to DNA ploidy or proliferative activity of colorectal carcinomas. So far, there are only a few studies relating DNA ploidy and CEA. Rognum et al. (8,9) and Scott et al. (10) reported that for aneuploid tumours plasma CEA values correlated with Dukes’ stage and also with the tumour content of CEA (9). Rognum et al. (11) have shown that an increase in CEA accompanies most recurrences of aneuploid
CEA, Ploidy, and Colorectal Carcinoma
tumours but only rarely recurrences of diploid tumours. In this study we have analysed prospectively the relationship between ploidy and preoperative CEA level in 130 patients with primary colorectal carcinoma and also in 30 patients with recurrence. SUBJECTS AND METHODS
Downloaded by [New York University] at 04:40 15 March 2016
Patients
One hundred and fifty-seven patients operated on at the Second Dept. of Surgery, Helsinki University Central Hospital, between November 1983 and April 1988 were included in this study. Patients with a previous history of cancer, hereditary nonpolyposis colorectal cancer syndrome, ulcerative colitis, or familial adenomatosis were excluded. The CEA value was measured at the time of diagnosis in 130 patients. In 27 patients preoperative CEA level was not measured for various reasons. Data on age, sex, and tumour stage, site, and size were collected from the patient records. In five cases the size of the tumour was not measured. The ACPS classification of tumour stage was used (12). The tumours were classified as right-sided (those proximal to splenic flexure), left-sided, and rectal. Follow-up
Follow up visits took place once every 6 months during the first 2 years, and thereafter yearly for up to 5 years. Routinely, the visits included a physical examination, sigmoidoscopy to exclude anastomotic recurrence, and blood tests including the measurement of CEA. Colonoscopy was done after 6 months and then 3 and 5 years after the primary operation. For the past 2 years the follow-up program has included the measurement of CEA every 3 months during the first 2 years. Recurrence was dated from the first evidence of relapse, based on physical, histologic, or imaging data. A recurrence was classified as local, localized distant, or disseminated. Tumour recurrence within the initial tumour bed and operative field, or anastomotic recurrence was defined as a local failure. Metastases amenable to complete surgical excision were defined as a localized distant failure. In 30 of the 36 patients with relapse the CEA
813
value was measured at the time of recurrence. The time of the death and the cause of death were determined from hospital records and from the Central Statistical Office of Finland. The patients have been followed up for a median of 36 months (range, 13 to 67 months). CEA determination The serum CEA was measured either with a double antibody assay (13) using commercially available CEA antisera (Dakopatts A/S, Copenhagen, Denmark) as first antibody or with the Abbot-CEA-RIA Diagnostic Kit (Abbot, Wiesbahn, Germany). The two assays showed a good correlation ( r 2 = 0.9997) in 100 samples with CEA concentrations ranging from normal up to 60,500 pg/l (14). In the current analysis a cut-off value of 5 pg/l was used. Flow cytometric analysis
Flow cytometric DNA analysis was performed as described previously (15). One to five biopsy specimens (0.5-1 cm3) were immediately frozen in liquid nitrogen and thereafter stored at -80°C until analysed. Necrotic areas were avoided. At the time of analysis the tumour samples were rapidly thawed in a 37 "C water bath and processed immediately into single-cell suspensions using scalpel and scissors. Chicken and trout erythrocytes were added as internal standards in most of the samples (16). The cell suspension was stained withethidium bromide (50 pg/mlin 10 mM Tris buffer, 1mMethylenediaminetetraaceticacid (EDTA), 0.3% Nonident P40, pH 7.5). The tube was vortexed and held on ice for 15 min. Then 0.25 ml of solution containing 1mg/ml RNAse (Sigma) was added to the tube, and the tube was incubated for 15min at room temperature. Immediately before analysis the sample was filtered through nylon mesh. For daily calibration HL-60 cells or monocytes were used. Fifty-eight tumours were analysed by a FACS IV and 72 by a EPICS C flow cytometer, With the FACS a 5-W and with the EPICS a 2-W argon-ion laser was used for excitation at 488 nm, 200mW, and the total emission above 590nm was measured. The mean coefficient of variation of the diploid peak was 4.9%.
M. Kouri et al.
814
Table I. The frequency of elevated CEA values related to clinicopathologic variables in 130 colorectal carcinoma patients
aneuploid G1/Go peak by the counts of the diploid GI/Go peak.
n
Percentage of patients with CEA 2 5.0 pg/l
32 41 57
38 44 39
69 61
46 33
42 33 55
38 39 42
Tumour stage Stage A Stage B Stage C Stage D Tumour size
22 53 29 26
32 30 28 81 *
5 3 . 5 cm 3.6-5.0 cm >5.0 crn
43 43 39
35 33 46
Statistical description and analysis Differences between frequencies were analysed with contingency tables. For correlation analysis Kendall’s tau test was used. For calculation of disease-free survival and overall survival, product limit survival analysis was performed, using the BMDP 1L computer program (18). Calculations of the significance of observed differences were performed with the log-rank test (Mantel-Cox). Patients with stage D disease were included in this material for the calculation of overall survival rates, but this category of patients was excluded for the calculation of the disease-free interval. The relative prognostic importance of all variables was investigated using Cox’s regression model and the BMDP 2L computer program (17). A prognostic variable with two or more categories of outcome is represented by a number of variables equal to the number of its categories minus one. The reference category was not included as a variable.
Variable Tumour site Right colon Left colon Rectum Sex
Downloaded by [New York University] at 04:40 15 March 2016
Male Female Age (years) 70
* p = 0.001,
The flow cytometric variables evaluated included the DNA ploidy and DNA Index (DI, where DI represented the ratio of the aneuploid stem line GI/GoDNA peak channel to the diploid stem line Gl/Go DNA peak channel). Tumours were classified as DNA-diploid and DNAaneuploid. and DNA-aneuploid tumours with a DI between 1.9 and 2.1 and a definable S- and G2/M-phase as DNA-tetraploid. The tumours were classified as aneuploid only if there were at least two G,/Go peaks. In tumours with multiple aneuploid peaks the DI of the aneuploid peak with the highest peak ratio was used. The S-phase fraction (SPF) was calculated on the basis of a rectangular model as described previously (17). The mean S-phase channel count was calculated from the shoulder of the G1/S-phase to the shoulder of the S/G2-phase. The proliferative index (PI) is the sum of the percentage of cells in G2/M-phase and S-phase. When multiple samples were analysed, the mean values of SPF and PI were used. SPF analysis was not possible in 31 cases. The peak ratio of aneuploid tumours was calculated by dividing the counts of the
Table 11. The frequency of elevated CEA values related to DNA ploidy, DNA index, SPF, PI, and peak ratio in 130 colorectal carcinoma patients Variable DNA ploidy Diploid Aneuploid Tetraploid DNA index
n
Percentage of patients with CEA 3 . 0 pg/l
41 75 14
34 48 14*
41 29 43 14 3
34 55 42 14 67
522% >22%
50 49
36 37
528% >28%
50 49
34 39
40 43
35 49
1 .o 1.1-1.4 1.5-1.8 1.9-2.1 2.2-3.1
SPF PI
Peak ratio 51.0 >l.O
*p
= 0.04.
CEA, Ploidy, and Colorectal Carcinoma
Downloaded by [New York University] at 04:40 15 March 2016
RESULTS CEA was increased ( 2 5 . 0 pg/l) in 40% (52 of 130) of patients before surgery, ranging from 5 to 28,900 pg/l (median, 17 pg/I). Eighteen patients had CEA values ranging from 5 to 10 pg/l. CEA was increased in 81% (21 of 26) of patients with stage-D disease, in contrast to 30% (31 of 104) of other patients ( p < 0.005) (Table 1). CEA was increased significantly more often in patients with aneuploid tumours than in those with diploid or tetraploid tumours (Table 11)for example, in 48% (36 of 75) of patients with aneuploid tumours, in 34% (14 of 41) of patients with diploid tumours, but in only 14% (2 of 14) of patients with tetraploid tumours ( p < 0.05). SPF, PI, or peak ratio was not associated with increased CEA values, although CEA was increased slightly more frequently in patients with DNA-aneuploid tumours with a higher peak ratio. The difference in CEA increase in patients with tumours of different ploidy pattern was also observed when analysed by stage (Fig. 1). CEA was not increased in any of the nine patients with stage-A or -B disease and a tetraploid tumour, but it was elevated in 38% (16 of 42) of the patients with aneuploid tumours and in 29% (7 of 24) of the patients with diploid tumours of the same stages. In stages C and D CEA was increased in 61% (20 of 33) of patients with aneuploid
Fig. 1. Percentage of patients with an increased CEA level at the time of diagnosis related to tumour stage and DNA ploidy ((0)= diploid, (a)= aneuploid, (a)= tetraploid).
1 1000
815 I
i
2 loo; 4
0
'
101
looo/
E3
1 StageA
Stage B
StageC
Stage D
Fig. 2. Box plot of increased CEA values ( 2 5 . 0 pg/l) in patients with colorectal carcinoma at the time of diagnosis. A = diploid tumours ( n = 14), B = aneuploid tumours (n = 36). CEA values were correlated to turnour stage in aneuploid turnours ( p < 0.005, Kendall rank test corrected for ties) but not in diploid turnours. The vertical lines indicate the range of 10% to 90% values; the boxes indicate 50% value; and the bars the median. Values outside the range are shown as separate points. There were two patients with diploid tumour of stage A and one with stage C, shown as a bar.
tumours, 41% (7 of 17) with diploid tumours, and 40% (2 of 5 ) with tetraploid tumours. The increased CEA values (25.0 yg/l) were associated with the stage in patients with aneuploid tumours ( p < 0.005, Kendall rank test) but not in patients with diploid tumours (Fig. 2). Thirty-six patients had a recurrence. Twenty patients had clinical symptoms, whereas 13 had only an increased CEA, 2 had an anastomotic recurrence observed on routine sigmoidoscopy, and 1 a perineal tumour. The median lead time was 4.6 months (range, 0-16 months) for the 13 patients with no clinical symptoms. A rise in CEA accompanied the recurrence in 70% (21 of 30) of all patients: in 77% (17 of 22) of the aneuploid, 40% (2 of 5 ) of the diploid, and 67% (2 of 3) of the tetraploid tumours (Table 111). CEA was negative more often in local recurrences than in others, but the difference did
816
M. Kouri et al.
Table 111. The frequency of elevated CEA values at the time of recurrence related to site of recurrence and DNA ploidy in 30 colorectal carcinoma patients Number of patients with CEA 25.0 pg/l Site of recurrence
Aneuploid
Tetraploid
Diploid
All
(%)
315 516 414
-
011 -
316 516 414
(50) (83) (100)
Local recurrence Localized distant recurrence Liver Lung
-
-
Disseminated recurrence Pelvis, abdomen, retroperitoneum
Liver Multiple sites
Downloaded by [New York University] at 04:40 15 March 2016
All
not reach statistical significance. In patients with an increased CEA level before surgery, a renewed CEA rise was observed in 91% (10 of 11) of the patients with recurrent disease but also in 60% (9 of 15) who had normal preoperative CEA values (NS). Nine of the 36 patients (25%) with recurrence were operated on radically; three of them had a local recurrence, four resectable liver metastases, and two resectable lung metastases. In all of these cases the primary tumour was aneuploid. In all but two of them the recurrence was accompanied by an elevated CEA level.
In patients with stage A-C disease ( n = 104) a preoperative CEA level 2 5 yg/l (n = 29) was associated with a worse prognosis. This difference was not statistically significant. Therefore, the level 210 yg/l ( n = 16) was also tested. With this cut-off point the difference between the two groups was statistically significant ( p < 0.05, logrank test) (Fig. 3). The CEA level was not an independent variable, since in a multivariate analysis with the Cox regression model only stage and DNA ploidy correlated independently with the disease-free survival. DISCUSSION
0 4 . 0
' 12
"
"
'
8
24 36 Months
48
CEA
ISSN: 0036-5521 (Print) 1502-7708 (Online) Journal homepage: http://www.tandfonline.com/loi/igas20
Serum Carcinoembryonic Antigen and DNA Ploidy in Colorectal Carcinoma M. Kouri, S. Pyrhönen, J.-P. Mecklin, H. Järvinen, A. Laasonen, K. Franssila, P. Kuusela & S. Nordling To cite this article: M. Kouri, S. Pyrhönen, J.-P. Mecklin, H. Järvinen, A. Laasonen, K. Franssila, P. Kuusela & S. Nordling (1991) Serum Carcinoembryonic Antigen and DNA Ploidy in Colorectal Carcinoma, Scandinavian Journal of Gastroenterology, 26:8, 812-818, DOI: 10.3109/00365529109037017 To link to this article: http://dx.doi.org/10.3109/00365529109037017
Published online: 08 Jul 2009.
Submit your article to this journal
Article views: 2
View related articles
Citing articles: 2 View citing articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=igas20 Download by: [New York University]
Date: 15 March 2016, At: 04:40
Serum Carcinoembryonic Antigen and DNA Ploidy in Colorectal Carcinoma A Prospective Study
Downloaded by [New York University] at 04:40 15 March 2016
M. KOURI, S. PYRHONEN, J.-P. MECKLIN, H. JARVINEN, A. LAASONEN, K. FRANSSILA, P. KUUSELA & S. NORDLING Dept. of Radiotherapy and Oncology and Second Dept. of Surgery, University Central Hospital; Pathology Laboratory of the Dept. of Radiotherapy and Oncology and Dept. of Bacteriology and Immunology, Helsinki University; and Dept. of Pathology, Helsinki University; Helsinki, Finland Kouri M, Pyrhonen S, Mecklin J-P, Jarvinen H, Laasonen A, Franssila K, Kuusela P, Nordling S. Serum carcinoembryonic antigen and DNA ploidy in colorectal carcinoma. A prospective study. Scand J Gastroenterol 1991, 26, 812-818 We have analysed the relationship between carcinoembryonic antigen (CEA) and DNA ploidy prospectively in 130 colorectal carcinoma patients. CEA was elevated preoperatively significantly more often in patients with DNA-aneuploid tumours than in DNA-diploid or DNA-tetraploid tumours-that is, in 48% (36 of 7 5 ) of patients with aneuploid tumours, in 34% (14 of 41) of patients with diploid tumours, but only in 14% (2 of 14) of patients with tetraploid tumours ( p < 0.05). Aneuploid tumours had an elevated CEA level in 38% of stage A-B disease and in 61% of stage C-D disease. The elevated CEA values (25.0 &I) correlated with tumour stage in patients with aneuploid tumours but not in patients with diploid tumours. Whereas CEA is a suitable marker for aneuploid carcinomas, other more sensitive tumour markers should be sought for diploid and also for tetraploid tumours. If such markers are found, flow cytometry could provide the most important information in selecting individual follow-up programmes for colorectal cancer patients. Key words: Carcinoembryonic antigen; colorectal carcinoma; DNA-ploidies; flow cytometry Mauri Kouri, M . D . , Dept. of Radiotherapy and Oncology, Helsinki University Central Hospital, SF-00290 Helsinki, Finland
Carcinoembryonic antigen (CEA), first described by Gold & Freedman (l),is clinically the most important marker of colorectal cancer. The serum CEA assay does not possess the sensitivity or the specificity required to distinguish between localized malignant tumours and benign disorders (2). In a recent report 26% of the patients with resectable primary colorectal carcinoma and 72% of the patients with unresectable or metastatic disease were CEA-positive when a cut-off value of 5.0 pg/1 was used (3). CEA is very useful in the postoperative monitoring of radically operated patients. The CEA assay can detect about 60% to 90% of recurrences in an asymptomatic stage regardless of the CEA value at the time of
operation of the primary tumour, and up to about half of such patients can be reoperated on with curative intent (4-7). On the other hand, all patients with recurrence do not show a CEA increase, for reasons that need further research. There is an increasing number of studies relating prognosis to DNA ploidy or proliferative activity of colorectal carcinomas. So far, there are only a few studies relating DNA ploidy and CEA. Rognum et al. (8,9) and Scott et al. (10) reported that for aneuploid tumours plasma CEA values correlated with Dukes’ stage and also with the tumour content of CEA (9). Rognum et al. (11) have shown that an increase in CEA accompanies most recurrences of aneuploid
CEA, Ploidy, and Colorectal Carcinoma
tumours but only rarely recurrences of diploid tumours. In this study we have analysed prospectively the relationship between ploidy and preoperative CEA level in 130 patients with primary colorectal carcinoma and also in 30 patients with recurrence. SUBJECTS AND METHODS
Downloaded by [New York University] at 04:40 15 March 2016
Patients
One hundred and fifty-seven patients operated on at the Second Dept. of Surgery, Helsinki University Central Hospital, between November 1983 and April 1988 were included in this study. Patients with a previous history of cancer, hereditary nonpolyposis colorectal cancer syndrome, ulcerative colitis, or familial adenomatosis were excluded. The CEA value was measured at the time of diagnosis in 130 patients. In 27 patients preoperative CEA level was not measured for various reasons. Data on age, sex, and tumour stage, site, and size were collected from the patient records. In five cases the size of the tumour was not measured. The ACPS classification of tumour stage was used (12). The tumours were classified as right-sided (those proximal to splenic flexure), left-sided, and rectal. Follow-up
Follow up visits took place once every 6 months during the first 2 years, and thereafter yearly for up to 5 years. Routinely, the visits included a physical examination, sigmoidoscopy to exclude anastomotic recurrence, and blood tests including the measurement of CEA. Colonoscopy was done after 6 months and then 3 and 5 years after the primary operation. For the past 2 years the follow-up program has included the measurement of CEA every 3 months during the first 2 years. Recurrence was dated from the first evidence of relapse, based on physical, histologic, or imaging data. A recurrence was classified as local, localized distant, or disseminated. Tumour recurrence within the initial tumour bed and operative field, or anastomotic recurrence was defined as a local failure. Metastases amenable to complete surgical excision were defined as a localized distant failure. In 30 of the 36 patients with relapse the CEA
813
value was measured at the time of recurrence. The time of the death and the cause of death were determined from hospital records and from the Central Statistical Office of Finland. The patients have been followed up for a median of 36 months (range, 13 to 67 months). CEA determination The serum CEA was measured either with a double antibody assay (13) using commercially available CEA antisera (Dakopatts A/S, Copenhagen, Denmark) as first antibody or with the Abbot-CEA-RIA Diagnostic Kit (Abbot, Wiesbahn, Germany). The two assays showed a good correlation ( r 2 = 0.9997) in 100 samples with CEA concentrations ranging from normal up to 60,500 pg/l (14). In the current analysis a cut-off value of 5 pg/l was used. Flow cytometric analysis
Flow cytometric DNA analysis was performed as described previously (15). One to five biopsy specimens (0.5-1 cm3) were immediately frozen in liquid nitrogen and thereafter stored at -80°C until analysed. Necrotic areas were avoided. At the time of analysis the tumour samples were rapidly thawed in a 37 "C water bath and processed immediately into single-cell suspensions using scalpel and scissors. Chicken and trout erythrocytes were added as internal standards in most of the samples (16). The cell suspension was stained withethidium bromide (50 pg/mlin 10 mM Tris buffer, 1mMethylenediaminetetraaceticacid (EDTA), 0.3% Nonident P40, pH 7.5). The tube was vortexed and held on ice for 15 min. Then 0.25 ml of solution containing 1mg/ml RNAse (Sigma) was added to the tube, and the tube was incubated for 15min at room temperature. Immediately before analysis the sample was filtered through nylon mesh. For daily calibration HL-60 cells or monocytes were used. Fifty-eight tumours were analysed by a FACS IV and 72 by a EPICS C flow cytometer, With the FACS a 5-W and with the EPICS a 2-W argon-ion laser was used for excitation at 488 nm, 200mW, and the total emission above 590nm was measured. The mean coefficient of variation of the diploid peak was 4.9%.
M. Kouri et al.
814
Table I. The frequency of elevated CEA values related to clinicopathologic variables in 130 colorectal carcinoma patients
aneuploid G1/Go peak by the counts of the diploid GI/Go peak.
n
Percentage of patients with CEA 2 5.0 pg/l
32 41 57
38 44 39
69 61
46 33
42 33 55
38 39 42
Tumour stage Stage A Stage B Stage C Stage D Tumour size
22 53 29 26
32 30 28 81 *
5 3 . 5 cm 3.6-5.0 cm >5.0 crn
43 43 39
35 33 46
Statistical description and analysis Differences between frequencies were analysed with contingency tables. For correlation analysis Kendall’s tau test was used. For calculation of disease-free survival and overall survival, product limit survival analysis was performed, using the BMDP 1L computer program (18). Calculations of the significance of observed differences were performed with the log-rank test (Mantel-Cox). Patients with stage D disease were included in this material for the calculation of overall survival rates, but this category of patients was excluded for the calculation of the disease-free interval. The relative prognostic importance of all variables was investigated using Cox’s regression model and the BMDP 2L computer program (17). A prognostic variable with two or more categories of outcome is represented by a number of variables equal to the number of its categories minus one. The reference category was not included as a variable.
Variable Tumour site Right colon Left colon Rectum Sex
Downloaded by [New York University] at 04:40 15 March 2016
Male Female Age (years) 70
* p = 0.001,
The flow cytometric variables evaluated included the DNA ploidy and DNA Index (DI, where DI represented the ratio of the aneuploid stem line GI/GoDNA peak channel to the diploid stem line Gl/Go DNA peak channel). Tumours were classified as DNA-diploid and DNAaneuploid. and DNA-aneuploid tumours with a DI between 1.9 and 2.1 and a definable S- and G2/M-phase as DNA-tetraploid. The tumours were classified as aneuploid only if there were at least two G,/Go peaks. In tumours with multiple aneuploid peaks the DI of the aneuploid peak with the highest peak ratio was used. The S-phase fraction (SPF) was calculated on the basis of a rectangular model as described previously (17). The mean S-phase channel count was calculated from the shoulder of the G1/S-phase to the shoulder of the S/G2-phase. The proliferative index (PI) is the sum of the percentage of cells in G2/M-phase and S-phase. When multiple samples were analysed, the mean values of SPF and PI were used. SPF analysis was not possible in 31 cases. The peak ratio of aneuploid tumours was calculated by dividing the counts of the
Table 11. The frequency of elevated CEA values related to DNA ploidy, DNA index, SPF, PI, and peak ratio in 130 colorectal carcinoma patients Variable DNA ploidy Diploid Aneuploid Tetraploid DNA index
n
Percentage of patients with CEA 3 . 0 pg/l
41 75 14
34 48 14*
41 29 43 14 3
34 55 42 14 67
522% >22%
50 49
36 37
528% >28%
50 49
34 39
40 43
35 49
1 .o 1.1-1.4 1.5-1.8 1.9-2.1 2.2-3.1
SPF PI
Peak ratio 51.0 >l.O
*p
= 0.04.
CEA, Ploidy, and Colorectal Carcinoma
Downloaded by [New York University] at 04:40 15 March 2016
RESULTS CEA was increased ( 2 5 . 0 pg/l) in 40% (52 of 130) of patients before surgery, ranging from 5 to 28,900 pg/l (median, 17 pg/I). Eighteen patients had CEA values ranging from 5 to 10 pg/l. CEA was increased in 81% (21 of 26) of patients with stage-D disease, in contrast to 30% (31 of 104) of other patients ( p < 0.005) (Table 1). CEA was increased significantly more often in patients with aneuploid tumours than in those with diploid or tetraploid tumours (Table 11)for example, in 48% (36 of 75) of patients with aneuploid tumours, in 34% (14 of 41) of patients with diploid tumours, but in only 14% (2 of 14) of patients with tetraploid tumours ( p < 0.05). SPF, PI, or peak ratio was not associated with increased CEA values, although CEA was increased slightly more frequently in patients with DNA-aneuploid tumours with a higher peak ratio. The difference in CEA increase in patients with tumours of different ploidy pattern was also observed when analysed by stage (Fig. 1). CEA was not increased in any of the nine patients with stage-A or -B disease and a tetraploid tumour, but it was elevated in 38% (16 of 42) of the patients with aneuploid tumours and in 29% (7 of 24) of the patients with diploid tumours of the same stages. In stages C and D CEA was increased in 61% (20 of 33) of patients with aneuploid
Fig. 1. Percentage of patients with an increased CEA level at the time of diagnosis related to tumour stage and DNA ploidy ((0)= diploid, (a)= aneuploid, (a)= tetraploid).
1 1000
815 I
i
2 loo; 4
0
'
101
looo/
E3
1 StageA
Stage B
StageC
Stage D
Fig. 2. Box plot of increased CEA values ( 2 5 . 0 pg/l) in patients with colorectal carcinoma at the time of diagnosis. A = diploid tumours ( n = 14), B = aneuploid tumours (n = 36). CEA values were correlated to turnour stage in aneuploid turnours ( p < 0.005, Kendall rank test corrected for ties) but not in diploid turnours. The vertical lines indicate the range of 10% to 90% values; the boxes indicate 50% value; and the bars the median. Values outside the range are shown as separate points. There were two patients with diploid tumour of stage A and one with stage C, shown as a bar.
tumours, 41% (7 of 17) with diploid tumours, and 40% (2 of 5 ) with tetraploid tumours. The increased CEA values (25.0 yg/l) were associated with the stage in patients with aneuploid tumours ( p < 0.005, Kendall rank test) but not in patients with diploid tumours (Fig. 2). Thirty-six patients had a recurrence. Twenty patients had clinical symptoms, whereas 13 had only an increased CEA, 2 had an anastomotic recurrence observed on routine sigmoidoscopy, and 1 a perineal tumour. The median lead time was 4.6 months (range, 0-16 months) for the 13 patients with no clinical symptoms. A rise in CEA accompanied the recurrence in 70% (21 of 30) of all patients: in 77% (17 of 22) of the aneuploid, 40% (2 of 5 ) of the diploid, and 67% (2 of 3) of the tetraploid tumours (Table 111). CEA was negative more often in local recurrences than in others, but the difference did
816
M. Kouri et al.
Table 111. The frequency of elevated CEA values at the time of recurrence related to site of recurrence and DNA ploidy in 30 colorectal carcinoma patients Number of patients with CEA 25.0 pg/l Site of recurrence
Aneuploid
Tetraploid
Diploid
All
(%)
315 516 414
-
011 -
316 516 414
(50) (83) (100)
Local recurrence Localized distant recurrence Liver Lung
-
-
Disseminated recurrence Pelvis, abdomen, retroperitoneum
Liver Multiple sites
Downloaded by [New York University] at 04:40 15 March 2016
All
not reach statistical significance. In patients with an increased CEA level before surgery, a renewed CEA rise was observed in 91% (10 of 11) of the patients with recurrent disease but also in 60% (9 of 15) who had normal preoperative CEA values (NS). Nine of the 36 patients (25%) with recurrence were operated on radically; three of them had a local recurrence, four resectable liver metastases, and two resectable lung metastases. In all of these cases the primary tumour was aneuploid. In all but two of them the recurrence was accompanied by an elevated CEA level.
In patients with stage A-C disease ( n = 104) a preoperative CEA level 2 5 yg/l (n = 29) was associated with a worse prognosis. This difference was not statistically significant. Therefore, the level 210 yg/l ( n = 16) was also tested. With this cut-off point the difference between the two groups was statistically significant ( p < 0.05, logrank test) (Fig. 3). The CEA level was not an independent variable, since in a multivariate analysis with the Cox regression model only stage and DNA ploidy correlated independently with the disease-free survival. DISCUSSION
0 4 . 0
' 12
"
"
'
8
24 36 Months
48
CEA
