Abstract
Previous randomized studies suggest that fecal occult blood test (FOBT) screening can reduce mortality from colorectal cancer (CRC). Our aim was to review the current status of FOBTs in CRC screening. FOB is measured using either the traditional guaiac-based tests or more recently introduced fecal immunochemical tests (FITs). FITs have several advantages over guaiac-based FOBTs, including higher sensitivity and specificity, resulting in improved clinical performance and higher efficiency. Another advantage in population screening according to European Guidelines for quality assurance in CRC screening is that FITs can be automated and user can adjust the cut-off at which a positive result is reported. In population-based screening, all those testing positively with any FOBT should be referred for colonoscopy. Conclusion: Although a plethora of FOBTs are available on the market, relatively few have been extensively tested for clinical sensitivity and specificity in CRC screening. Current data imply that new FITs have superior test characteristics as compared with guaiac-based FOBTs. The latest development in the field is represented by the proteomic-based tests that may further reduce false-negative rates in CRC screening. Simple stool sample preservation and automatic analysis are other important issues in population-based screening for CRC.
- Colorectal cancer
- diagnosis
- screening
- fecal occult blood test
- FOBT
- fecal immunochemical test
- FIT
- false-negative
- false-positive
- review
Colorectal cancer (CRC) is the third most common cancer worldwide, with over 1.3 million new cases and over 600,000 deaths each year (1). In Finland, the incidence of CRC is lower than in many Western countries but mortality is quite similar, reflecting a mortality-to-incidence ratio disparity in Finland. During the 3-year period of 2015 to 2017, incident CRC was encountered in 4,577 women and in 5,131 in men, with an estimated annual average of 3,236 new CRC cases (2). After lung cancer, CRC is the second most common cause of cancer deaths among men and women combined in Finland (2).
The majority of CRCs develop from adenomas or adenomatous polyps and several studies have shown the efficacy of screening for detection of large adenomas. CRC screening can achieve the goals of both primary prevention (by detecting cancer precursors: polyps, adenomas) and secondary prevention (by detecting early cancers) (3, 4). Thus, organized CRC screening offers a possibility for cancer prevention and early detection of cancer, with reduced mortality (5, 6).
Although a plethora of fecal occult blood test (FOBT) are available on the market, relatively few of them have been extensively tested for clinical sensitivity and specificity in CRC screening. The main aim for the use of FOBTs is to reduce CRC mortality (7-11). European Guidelines for quality assurance in CRC screening recommend fecal immunochemical tests (FITs) because FITs have improved test characteristics compared to guaiac-based FOBTs (gFOBTs) (12) (Table I).
The most commonly used CRC screening tests include flexible sigmoidoscopy, colonoscopy and FOBT (13, 14). Lin et al. reviewed literature and found four randomized clinical trials (RCTs) evaluating one or two rounds of flexible sigmoidoscopy (n=458,002), showing reduced CRC-specific mortality compared with no screening [incidence rate ratio=0.73; 95% confidence intervaI (CI)=0.66-0.82]. In addition, five RCTs with multiple rounds of biennial screening with gFOBTs (n=419,966) showed reduced CRC-specific mortality [from relative risk (RR)=0.91, 95% CI=0.84-0.98 at 19.5 years to RR=0.78, 95% CI=0.65-0.93 at 30 years from screening] (15).
gFOBTs have been the tests most often used in organized, population-based screening programs (16-18), and RCTs show that screening by FOBT can reduce CRC mortality by 18-33% (7, 9, 19). Two main types of FOBT exist: gFOBTs and FITs. The gFOBT uses the pseudoperoxidase activity of intact or free hemoglobin, based on the oxidation of guaiac by hydrogen peroxidase, and reacts with any peroxidase in feces. Therefore, the gFOBT test is not fully specific for human hemoglobin and the reaction is complicated by reaction of several foods with any peroxidase content, certain chemicals or medications (20, 21). On the other hand, the FIT detects early degradation products and the globin moiety of intact human hemoglobin (22-24). The gFOBT was the first technology on the market, and its clinical efficacy has been investigated more extensively than that of FITs (22-25). Because gFOBTs are not specific for human blood, it is necessary to confirm a FOBT-positive result by colonoscopy (26, 27).
Our aim in this article is to: i) Critically review the known limitations of gFOBT and FIT in CRC screening; and ii) introduce a modern genomics- and proteomics-based strategy for CRC screening.
gFOBT
Different gFOBTs have become very popular screening methods for CRC (15). The gFOBT has been on the market for several decades, and the test is quite easy to perform and inexpensive. In FOBTs, multiple samples are needed to detect CRC with high sensitivity. These tests detect FOB based on the peroxidase activity of hemoglobin-derived heme groups but, unfortunately, this reaction is not specific for human blood. In addition to human blood, these guaiac-based tests can also trace animal blood derived from food, and in addition, peroxidases derived from some raw vegetables. This can lead to false-positive results and unnecessary referrals for colonoscopy. In addition, these tests are not highly sensitive, which can also lead to false-negative results (15).
Several gFOBTs are available on the market with different sensitivity and specificity in detecting CRC (28, 29). Rabeneck et al. reviewed repeated annual or biennial gFOBT testing, showing gFOBT sensitivity for CRC from 51% to 100% and specificity from 90% to 97%, with a positive predictive value of between 2.4% and 17.0% (28).
Although screening by gFOBTs has been shown to reduce mortality from CRC, the data on the efficacy of gFOBT screening are controversial (15). Hewitson et al. reviewed four RCTs, showing a 16% reduction in the RR of CRC mortality in the screened population (RR=0.84, 95% CI=0.78-0.92) (29). In the Minnesota trial, the RR was adjusted for attendance at screening, the overall predicted reduction in relative mortality was 25% amongst those screened (RR=0.75, 95% CI=0.66-0.84) (7).
Although previous RCTs used the Hemoccult II test (Beckman, Fullerton, CA, USA), several non-randomized studies (30-33) used different gFOBTs, including Hema-screen (Immunostics, Ocean, NJ, USA). In conclusion, when critically assessed in a recent systematic review (15), a meta-analysis of all four FOBT screening trials indicated no benefit for all-cause mortality (RR=1.00, 95% CI=0.99-1.03) (15, 22). This made the authors suggest that it is not be expected that CRC screening would reduce all-cause mortality in these ongoing FOBT trials (15). In addition, because of their poor sensitivity and low specificity, gFOBTs have become increasingly replaced by FITs in population-based CRC screening programs (34, 35).
FITs
Since the invention of the immunochemical test principle by Suovaniemi et al. in the 1980s (36), an increasing number of FITs have been developed, particularly in Japan, the pioneering country of CRC screening, where different FITs have been the principal screening method since the early 1990s (37). The FIT is based on the detection of the globin moiety of human hemoglobin or its degradation products. In a recent systematic review of FITs in CRC screening (21, 27), 12 types of FITs were identified in the literature, representing 20 different proprietary names. Due to the major differences in test methodology, the authors were unable to assess all FITs as a class, however, and undertook a sub-group analysis (15, 21). On the basis of a single stool specimen, the most commonly evaluated FITs demonstrated good sensitivity (range=73%-88%) and specificity (range=90%-96%). Lin et al. reviewed one study (n=9,989), reporting that FIT plus stool-based DNA test had better sensitivity (92%) but lower specificity (84%) in detecting CRC than FIT alone (15).
The only RCT comparing gFOBT and FIT published so far of van Rossum et al. concluded that the performance of FITs is clearly superior to that of gFOBTs in detecting any type of colorectal neoplasia (38). This has led to the rapid emergence of a large number of commercial FIT products onto the market. Lee et al. reviewed the accuracy of FITs for CRC screening and, after careful selection, they found 18 out of the 53 available studies to be eligible for their formal meta-analysis. These studies included eight different commercial FITs, but not the ColonView quick test (Biohit Oyj, Helsinki, Finland) because it was not yet on the market. In this meta-analysis, the pooled sensitivity and specificity for the CRC endpoint were: 79% (95% CI=69-86%) and 94% (95% CI=93-97%), respectively (39). These meta-analytical results are supplemented by two separate studies testing ColonView in head-to-head comparison with gFOBT (Hemoccult SENSA; Beckman Coulter, Pasadena, Los Angeles, CA, USA) (26, 27). In a St. Petersburg study of a cohort of 300 patients referred for colonoscopy, the ColonView test showed pooled sensitivity of 100% and specificity of 95% for proximal colon neoplasia as well as 98% sensitivity and 95% specificity for distal colon neoplasia (26). Guimarães et al. conducted a clinical trial comparing the ColonView test with gFOBT in a similar setting of 368 colonoscopy-referral patients at Barretos Cancer Hospital in Brazil. For the CRC endpoint, the ColonView test had 95% sensitivity and 65% specificity [area under the receiver operating characteristics curve (AUC)=0.799], while the gFOBT had sensitivity of 76% and specificity of 84% (AUC=0.800). For the adenoma endpoint, the difference in sensitivity between ColonView test and gFOBT was even larger (sensitivity of 44% versus 19%, respectively). The authors concluded that due to its 95% sensitivity, ColonView test is superior to gFOBT in organized CRC screening (27).
Genomics and Proteomics-based Approaches in CRC Screening
Compared with gFOBTs and iFOBTs, the use of genomics- and proteomics-based approaches in population-based CRC screening has been less widely investigated. The aim in using fecal DNA markers in CRC screening is to find mutant DNA present in stool. Imperiale et al. investigated a DNA panel of 21 separate mutations in several tumor protein genes, as well as the detection of microsatellite markers. They concluded that the DNA marker panel displayed a higher sensitivity than the gFOBT, without a marked reduction in specificity. However, the gFOBT result was based on only a test from a single time point (40).
Ahlquist et al. investigated the same DNA panel with two different gFOBTs in a multi-center study of 2,497 asymptomatic individuals; the sensitivity for CRC was 20% for the DNA test, 11% for Hemoccult II and 21% for Hemoccult SENSA specificity was 96% for the DNA panel, 98% for Hemoccult and 97% for Hemoccult SENSA (41).
RNA-based CRC screening methods may offer a more promising strategy than use of fecal DNA markers alone (42-46). A microRNA strategy might meet the criteria of an optimal CRC screening test: It is non-invasive; 1 g of stool is adequate for testing; sampling on consecutive dates is not required; stool samples can be sent by mail; the test is able to differentiate between normal tissue and CRC; and the test can be automated (47). In their first article, Ahmed et al. addressed the experimental design and selected 10 colon cancer genes to monitor changes at various stages in the neoplastic process particularly applicable for screening of early-stage CRC. Although some of the genes in CRC tissue showed less variability, stool, however, was suitable for CRC screening and the so-called transcriptomic molecular strategy using tissue or stool samples offerred higher sensitivity and specificity than currently used DNA markers (42). In their second article, the same authors described the standardization strategy and test performance of transcriptomic molecular markers in CRC screening. The preservation of stool samples prior to RNA extraction is important, and use of an appropriate preservative and keeping the stool at 4°C during transport is recommended (43). Total RNA extraction kits are commercially available, containing buffer which removes bacterial RNA from stool samples, leaving the undegraded RNA of human origin.
In order to determine the sensitivity and specificity of transcriptomic CRC screening, it is important to conduct prospective RCTs. In 2014, in studying 41 patients with CRC and 54 healthy controls, Koga et al. demonstrated that a highly sensitive DNA chip assay had higher sensitivity and specificity in detecting early-stage CRC (48).
Since sporadic CRC is one of the most frequent types of cancer in Western world, a large number of other biomarkers and analyses are available to predict disease prognosis and to help in CRC screening (46, 49-54). Auge et al. used a compact fully-automated immunochemistry analyzer (Kroma It; Linear Chemicals S. L. Spain, distributed in Spain by Laboratorios LETI, S. L. Unipersonal, Spain) for fecal occult hemoglobin, with sensitivity and specificity for CRC of 36% and 92%, respectively (49). Christensen et al. (50) determined plasma TIMP metallopeptidase inhibitor 1 (TIMP1) and carcinoembryonic antigen (CEA) as markers for CRC using an automated analysis platform. Due to the small number of patients with CRC (n=32), sensitivity and specificity were not reported but the AUC was 0.731 for CEA, 0.695 for TIMP1, and 0.753 for CEA combined with TIMP1. Dressen et al. evaluated the diagnostic performance of a new magnetic multiplex immunoassay including several biomarkers for CRC diagnosis. CEA showed the best performance, with an AUC of 0.859. A combination of CEA and cancer antigen 19-9 had a higher AUC (0.893) as compared to either biomarker alone. They concluded that CRC diagnosis could be improved by a new biomarker classes and their combination by novel multiplex immunoassay (51). Bruns-Toepler et al. evaluated a new stool sample collection device with increased buffer stability for FIT, being a particularly promising tool for large-scale screening of CRC due to its advanced properties in sample handling, stability and hemaglobin analysis (52). Venäläinen et al. used urinary metabolomics to identify a panel of polyamine profiles for detecting CRC. They concluded that the determination of urinary polyamines by liquid chromatography–mass spectrometry can be used to differentiate those with CRC from healthy individuals (53).
Conclusions and Relevance – Which Fecal Test Should Be Chosen in CRC Screening?
There is little doubt that the characteristics of the FITs are superior to those of gFOBT as the screening tool for CRC (12; Table I). The WHO Guidelines recommend a panel of gFOBT and FIT (22). This strategy has been studied using an approach in which those with a weakly positive or equivocal result on gFOBT were asked to complete a FIT with a tube and a card collection device (55, 56). This strategy may reduce the number of false-positive results in CRC screening.
European Guidelines for quality assurance in CRC screening recommend FIT tests because FITs have improved test characteristics compared to gFOBTs (12). FITs have higher sensitivity and specificity, can be automated and the user can adjust the cut-off at which a positive result is reported. In conclusion, FITs are currently the test of choice for population-based CRC screening (Table I).
The latest development in the field of CRC screening is represented by genomics- and proteomics-based approaches, both of which have been less intensely studied. The strategy of using fecal DNA markers in CRC screening is based on the finding that mutant DNA is excreted in stool. At present, RNA-based methods look more promising than the use of fecal DNA markers only. Of the latter, a miRNA strategy seems to be among the most optimal for CRC in the future.
Acknowledgements
The study was funded by the Heikki, Aino and Aarne Korhonen foundation and the EVO-funding of the Kuopio University Hospital, Finland.
Footnotes
Authors' Contributions
All Authors have met all of the following four criteria: i). Substantial contributions to the conception or design of the work or the acquisition, analysis, or interpretation of data for the work. ii). Drafting the work or revising it critically for important intellectual content. iii). Final approval of the version to be published. iv). Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
This article is freely accessible online.
Conflicts of Interest
The Authors report no conflicts of interest or financial ties to disclose. The Authors alone are responsible for the content and writing of this article.
- Received January 8, 2020.
- Revision received January 15, 2020.
- Accepted January 16, 2020.
- Copyright© 2020, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved