Assessment of Serum CEA Levels in Different Radiological stages and Histopathological Grades of Colorectal Malignancy

Abstract

The study assessed the usefulness of serum CEA in predicting radiological stages and histopathological grades of colorectal malignancy in 202 patients admitted with radiologically and histopathologically proven colorectal malignancy. Histopathology, 68.96% were well differentiated; 42.55% moderately differentiated and 60% poorly differentiated adenocarcinoma. Patients had high serum CEA levels (>5ng/m1). On analyzing radiological staging, 72.7% with high serum CEA (> 5ng/ml) were node positive [stage III & 1\7]; 68.47% patients with low serum CEA (<5ng/m1) were node positive. Considering serum CEA, 80.2% patients had serum CEA value more than 2.5ng/ml and 54.45% patients had elevated serum CEA value more than 5ng/ml. It is concluded that serum CEA has got statistically little value in predicting histopathological grading and staging of colorectal malignancy.

Keywords

Colorectal malignancy, Radiological stage, Histopathological grade, serum CEA

Introduction

Adenocarcinoma of the colon and rectum is the third most common site of new cancer cases and death in men and women. The American cancer society suggest that one in twenty people will develop colorectal cancer in their life time risk being slightly higher for man than women.

The carcino embryogenic antigen (CEA), an onco fetoprotein was originally reported by Gold and Freedman in 19651. Quantitative estimation is done by using direct radioimmunoassay. S. CEA level >2.5ng/m1 is significant, said to be elevated > 5 ng/m15. In 1974, FDA approved marketing of the CEA test for the indications viz. early detection , diagnosis, screening of patients at high risk to develop cancer e.g. ulcerative colitis and chronic liver disease, and monitoring treated patients for cancer recurrence , the effects of radiation and chemotherapy4. Recently it is suggested that of a prognostic indicator of resect able colorectal malignancy further studies have put the true valve of CEA test for the above mentioned indications in a questionable state.

Staging of Colorectal Malignancy
Tx: No description of the tumor’s extent is possible because of incomplete information.

Grading of Colorectal Malignancy
Grading is based on worst area. It is acceptable to grade simply as Low vs. High grade. It applies only to adenocarcinoma NOS Low grade ≥50% gland forming

Well differentiated – >95% gland forming Moderately differentiated – 50- 95% gland forming High grade <50% gland forming

Poorly differentiated – 0-49% gland forming
Staging of Colorectal Malignancy
Tx: No description of the tumor’s extent is possible because of
incomplete information.

Tis: The cancer is in the earliest stage (in situ). It is only in the mucosa and has not grown beyond the muscularis mucosa (thin inner muscle layer).

T1: The tumor has grown through the muscularis mucosa and extends into the submucosa.

T2: The tumor has grown through the submucosa and extends into the muscularis propria (thick outer muscle layer).

T3: The tumor has grown through the muscularis propria and into the outermost layers of the colon or rectum but not through them. It has not reached any nearby organs or tissues.

T4a: The cancer has grown through the serosa (also known as the visceral peritoneum), the outermost lining of the intestines.

T4b: The cancer has grown through the wall of the colon or rectum and is attached to or invades into nearby tissues or organs.

N categories for colorectal cancer indicate if the cancer has spread to nearby lymph nodes and, if so, how many lymph nodes are involved. To get an accurate idea about lymph node involvement, most doctors recommend that at least 12 lymph nodes be removed during surgery and looked at under a microscope.

Nx: No description of lymph node involvement is possible because of incomplete information.

N0: No cancer in nearby lymph nodes.

N1: Cancer cells are found in or near 1 to 3 nearby lymph nodes

N1a: Cancer cells are found in 1 nearby lymph node.

N1b: Cancer cells are found in 2 to 3 nearby lymph nodes.

N1c: Small deposits of cancer cells are found in areas of fat near lymph nodes, but not in the lymph nodes themselves.

N2: Cancer cells are found in 4 or more nearby lymph nodes

N2a: Cancer cells are found in 4 to 6 nearby lymph nodes.

N2b: Cancer cells are found in 7 or more nearby lymph nodes.

M categories for colorectal cancer indicate whether or not the cancer has spread (metastasized) to distant organs, such as the liver, lungs, or distant lymph nodes.

M0: No distant spread is seen.

M1a: The cancer has spread to 1 distant organ or set of distant lymph nodes.

M1b: The cancer has spread to more than 1 distant organ or set of distant lymph nodes, or it has spread to distant parts of the peritoneum (the lining of the abdominal cavity).

Grading of Colorectal Malignancy
Grading is based on worst area. It is acceptable to grade simply as Low vs. High grade. It applies only to adenocarcinoma NOS Low grade ≥50% gland forming

Fig. 1: Age distribution of patients with colo-rectal cancer
Fig. 2: Gender distribution of patients with colo-rectal cancer
Fig. 3: Exposure to smoking among patients with colorectal cancer. 105 – nonsmokers and 97 – smokers
Fig. 4: Serum CEA level in patients with colo-rectal cancer at 2.5ng/ml level
Fig. 5: Serum CEA level in patients with colo-rectal cancer more than 5ng/ml level.
Fig. 6: Histopathological grading of colo-rectal cancer.
Fig. 7: CT staging of colo-rectal cancer
Fig. 8: Node positivity in cases of colo-rectal cancer 143 node positive cases and 59 node

Well differentiated – >95% gland forming Moderately differentiated – 50-95% gland forming High grade <50% gland forming

Poorly differentiated – 0-49% gland forming

Materials and Methods

202 patients admitted in surgical wards of Govt. Medical College, Thiruvananthapuram with radiologically and histopathologically proven colorectal malignancy were studied for a period of around 22 months. The study was based on clinical records. Those patients who were previously diagnosed with adenocarcinoma other than CRM (colorectal malignancies) were excluded. For those who satisfied the inclusion criteria, non-probability sampling technique was applied; data regarding the
histopathological grade, radiological stage based on contrast enhanced CT scan and serum CEA valve was collected; we proceeded with diagnostic test evaluation. Statistical analysis was carried out using SPSS 16.0 software (SPSS.Inc, Chicago, USA). ANOVA test was applied for correlating serum CEA and histopathological grade, same was done for correlating serum CEA and CT stage, t test was applied for correlating staging based on node positivity. Logistic regression was done to find the predictive value of serum CEA. ROC curve was used to find out the cut off value of CEA

Results

Out of 202 patients under the study, mean (SD) age was 60 (10.67) years. Minimum age was 20 years and maximum 92 years (Fig. 1). Among them, 45.5% were females and 54.5% were males (Fig. 2).105 nonsmokers (52%) and 97 nonsmokers (48%) were included (Fig. 3). Serum CEA value was found to be elevated > 2.5 ng/ml in 80.2 % patients (Fig. 4) and to be elevated >5 ng/ml in 54.45% (Fig. 5). Histopathologically the lesions were well differentiated in 58, moderately differentiated in 94 and poorly differentiated in 50 (Fig. 6). The TNM staging of the cases according to CT scan findings is given in Fig. 7. Histopathologically, 143 cases were node positive and 59 node negative (Fig. 8).

Statistical Study

On applying chi square test, 72.7% cases showed node positivity in patients with S.CEA value more than 5 ng/ml i.e 80 out of 110 cases. But the test was shown to be statistically not significant, as p value was 0.98 (Table 1)

Table 1. Chi square test comparing node positivity and serum CEA values.80 out of 110 cases,72.7% cases showed node positivity in patients with S.CEA value more than 5 ng/ml.30 node negative patients were also having more than 5ng/ml serum CEA.

P value – 0.98

ANOVA TEST was applied for correlating S.CEA and histopathological grading. Here the median CEA was more for well differentiated malignancy i.e 7.7 with IQR (InterQuartile Range) of 31.1 and for moderately differentiated and poorly differentiated CRM, 4.2 & 6.5 respectively. But as p value was 0.279, statistically it was not significant (Table 2 and Fig. 9). IQR is more for well differentiated malignancy which means the values have got wide range of dispersion.It is shown in box plot. Table 2. ANOVA test comparing serum CEA values and histopathological grading.

P value – 0.279
Fig. 9: ANOVA test comparing serum CEA values and histopathological grading.

ANOVA was done for CT staging and S.CEA. Median was gradually increasing but got deviated for the stage IIc (42.4 IQR means wide range of dispersion). p value was 0.55 hence statistically insignificant correlation (Table 3 and Fig. 10).

Table 3. ANOVA test comparing serum CEA values and TNM staging.

P value – 0.55
Fig. 11: ANOVA test comparing serum CEA values and TNM staging

t-test as applied for correlating S.CEA and staging based on node positivity with median 5.1 and 6 for node negative and node positive malignancies respectively i.e with slight difference.(p value 0.59) This is shown in Table 4. Table 4. ANOVA test comparing serum CEA values and TNM staging.

Logistic regression was done to find the predictive value of S.CEA. Though serum CEA level as found to be a significant predictor for histopatological grading (p-0.048) , COX and SNELL’S R2 value was 0.072(7.2%) which means only 7.2 % of variability could be predicted.

ROC Curve

As the area under the curve (ROC) was found to be low, Serum CEA has got statistically little value in predicting histopathological grading and staging of colorectal malignancy (Fig. 12).

Fig.12.ROC curve to test Sensitivity and Specificity.

Youden index:used for assessing the maximum potential effectiveness of a biomarker, is a common measure of ROC curve. For a diagnostic test to be predictive,its sensitivity and specificity should be maximum as well as youden’s index should be nearing to 1.here we got 48.25 sensitivity, 64.41
specificity and 0.1266 Youden’s index

Table 5. Youden index for comparing sensitivity and specificity

Discussion

Our primary intention for conducting the study was that pre-operative serum CEA value would help in predicting radiological stage and histopathological grade of colorectal malignancies and hence will contribute to the prognosis also. On doing logistic regression, serum CEA level was claimed to be a significant predictor for histopathological grading (p-0.048), poor differentiation in colon cancers is more likely to be associated with reduced ability to produce this antigen. Still Cox and Snells R2 value was 0.072 (7.2%) which means only 7.2% variability could be predicted. In this study, it is difficult to proclaim that pre-operative CEA got any predictive value for radiological staging and histopathological grading of CRM.

Steele et al studied 316 patients for a median time of 3 years and 9 months followed upto 8 months using 5nanogram per milliliter as an end point they found pre-operative CEA test to be unrelated to prognosis in patients with CRM, they only found significant correlation with recurrence free survival in colon cancer patients with 1 to 4 nodes involved2.

Moertel et al in 1986 did show a significant co relation between CEA and prognosis but this coreleation was weakned, when corrected for stage more than comparing for the nodal status1. Goslin et al reported that CEA is of no prognostic value in patients with stage B lesions6, as well they extended it to include stage C. this conclusion stood still regardless of whether tumours are well differentiated, moderately differentiated or poorly differentiated3

Conclusion

The study aimed at establishing the usefulness of preoperative serum CEA assessment in predicting the radiological stages and histopathological grades of colorectal malignancies. 202 patients with proven CRM were studied for 22 months. On diagnostic test evaluation, the area under the curve (ROC) was found to be low; hence we reached the conclusion that serum CEA has got statistically little value in predicting histopathological grading and staging of colorectal malignancies.

References

  1. Moertel CG et al, Carcinoembryonic Antigen Test in the Diagnosis, Staging, and Prognosis of Colorectal Cancer. Mayo Clinic and Mayo Foundation, Rochester published in 1986()
  2. Steele G Jr, Ellenberg S, Ramming, K et al. CEA monitoring among patients in multiinstitutional therapy protocols. Ann Surg 1982
  3. Goslin R, O’Brien MJ, Steele G et al. correlation of plasma CEA and CEA tissue staining in poorly differentiated colorectal malignancies Cancer 58:603- 610,1986
  4. Gold P, Freedman SO. Specific carcino-embryonic antigen of the human digestive system. JExp Med 1965; 122:467-481.
  5. Sabiston. Textbook of Surgery 1st south Asia edition; Tumour markers vol1 page698
  6. Goslin R, Steele G Jr, Mac Intyre J et al. the use of preoperative plasma CEA levels for the stratification of patients after curative resection of colorectal cancers Ann surg 1980;192:747-751

Acknowledgements: None.
Conflict of Interest: None
Address for Correspondence:
Dr. Greeshma S,
Junior resident, Department of General Surgery,
Govt. Medical College, Thiruvananthapuram 695011