Sensitivity and Specificity of a Quantitative Immunochemical FOBT for Colorectal Neoplasia

David A. Johnson, MD, FACG, FACP


April 10, 2007

A Quantitative Immunochemical Fecal Occult Blood Test for Colorectal Neoplasia

Levi Z, Rozen P, Hazazi R, et al
Ann Intern Med. 2007;146:244-255


The screening rates for colorectal cancer, although increasing as a result of enhanced insurance coverage and patient education, remain low relative to other well-established screening programs. It has been shown that stool-based testing for occult blood reduces colorectal cancer incidence and related mortality. This type of testing relies on the guaiac-based detection of the peroxidase activity of hemoglobin. The drawbacks of this testing strategy include a relatively poor sensitivity for colorectal cancer (even poorer for precancerous adenomas) and a poor specificity. Therefore, it is associated with a high rate of false-positive results. Additionally, the test is not specific for human hemoglobin and thus, food containing hemoglobin (red meats) or other foods with pseudoperoxidase activity may yield positive results.

Immunochemical fecal occult blood tests (I-FOBT) were developed with the goal of improving specificity and eliminating the cross-reactivity for dietary influences of hemoccult-based FOBT. I-FOBT involves the use of one or more monoclonal or polyclonal antibodies to detect human hemoglobin. This test, which reports findings qualitatively (ie, as positive or negative), has been recommended as a replacement for FOBT by the American Cancer Society, although it has not yet been recommended as such by the Multisociety Colorectal Task Force Guidelines group nor has it become an established clinical practice in the United States. One unique aspect of the I-FOBT is that it allows for variation of the threshold for detection of human hemoglobin and therefore may offer varying ranges of sensitivity and specificity.

This prospective cross-sectional study involved 1000 consecutive ambulatory patients, some of whom were symptomatic, who were undergoing elective colonoscopy in Israel. The hemoglobin content of 3 bowel movements was measured and the highest value of these 3 stools was taken as the reportable result and compared with the colonoscopic findings. Forty-nine patients were excluded because of incomplete colonoscopy. Polyps were removed and analyzed by pathologists blinded to the I-FOBT results. Clinically significant neoplasia was defined as colon cancer or advanced polyps (≥ 1 cm or with ≥ 20% villous histology, or any high-grade dysplasia).

The patients submitted I-FOBT results from 3 daily or consecutive stool samples before the colonoscopy. There were no restrictions on diet or medications other than discontinuing aspirin and anticoagulant therapy, when present. The diagnostic value of the I-FOBT was assessed using thresholds for detection at the manufacturer's suggested threshold for detection of 100 ng/mL of buffer, as well as at different thresholds ranging from 50 ng/mL to 200 ng/mL. At the 100-ng/mL fecal hemoglobin threshold, the sensitivity and specificity for detection of colorectal cancer was 88% and 90%, respectively, and for detection of "significant neoplasia" was 62% and 93%, respectively. Variation of the threshold to 75 ng/mL of buffer improved the sensitivity for detection of cancer and significant neoplasia to 94% and 67%, respectively, but at the trade-off of a reduced specificity of 88% and 91%, respectively.

The correlation coefficients between any two of the 3 I-FOBT samples showed a range of 0.553 to 0.599. These moderate correlations presumably reflect the variability in day-to-day blood loss from the colon neoplasm and are analogous to the variability seen in testing with the guaiac-based FOBT. To eliminate work-up bias in those patients who were symptomatic or had a positive FOBT, the cancers discovered in these individuals was removed from the analysis. This lowered the cancer detection sensitivity slightly from 94% to 90% but removed patients that might be deemed "higher risk" from the true "average-risk" population.


Although this study was not conducted in a pure "screening" population, the removal of patients undergoing directed testing because of clinical reasons makes the remaining population more reflective of true screening. Despite this adjustment, however, there was no significant loss in the sensitivity -- which was remarkably high for colorectal cancer compared with previous FOBT reports.

The unique feature of having variable thresholds to choose from may allow the physician to adjust the thresholds to suit the patient's clinical characteristics. For example, if an elderly patient with multiple medical problems were being considered for screening but there was some hesitancy regarding performing a potentially unnecessary colonoscopy, the threshold could be set to maximize specificity, albeit at a recognizable potential loss in specificity. Although we are not yet at this point, it is conceivable that these types of discussions between patients and their care providers may occur in the future. For example: "With this threshold for blood detection, you have an (X) % chance of cancer or advanced adenoma." Another advantage of I-FOBT is that this test is automated, precluding the variability of test interpretation that is evident with FOBT.

Before one might think this new test to be "perfect," further validation of these high rates of success is clearly warranted. The question remains in this physician's mind as to how such high rates are achieved, as bleeding from cancers, much less polyps, is quite sporadic and variable -- thus accounting for the lower success rates achieved with fecal blood detection tests to date.

A recent study from France[1] compared the performance of FOBT and I-FOBT among 10,673 average-risk patients. These authors found that at a threshold of 75 ng/mL, there was a slight gain in the positive predictive value for cancer detection with I-FOBT vs FOBT (8.7 vs 7.3) but an even greater gain in the positive predictive value for advanced adenomas (49.2 vs 27.7, respectively). However, not every patient with a positive test underwent colonoscopy. Therefore, although I-FOBT may appear to be "better" than FOBT, it will be up to the insurers to pay for this more expensive test (cost: $18-$30; Medicare reimbursement: $22 compared with standard FOBT [cost: $1-$3; Medicare reimbursement $4.72]) and it will be up to guideline committees to consider how this test will be profiled in the growing "menu" of nonendoscopic screening options for colorectal cancer.

Abstract URL:


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.