The rate of positive tests using fecal immunochemical test (FIT) does not decrease with subsequent campaigns, but the positive predictive value of advanced neoplasia significantly decreases in subsequent campaign after a first negative test. A relationship between the fecal hemoglobin concentration (Fhb) and the opportunity to detect a colorectal cancer in subsequent campaign has been shown.

To predict the severity of colorectal lesions based on Fhb measured during previous colorectal cancer screening campaign.

This etiological study included 293750 patients aged 50-74, living in Auvergne-Rhône-Alpes (France). These patients completed at least two FIT [test_(-1) and test₍₀₎] between June 2015 and December 2019. Delay between test_(-1) and test₍₀₎ was > 1 year and test_(-1) result was negative (< 150 ngHb/mL). The severity of colorectal lesions diagnosed at test₍₀₎ was described according to Fhb measured at test_(-1) [Fhb_(-1)]. The relationship between the severity classified in seven ordinal categories and the predictive factors was analyzed in an ordered multivariate polytomous regression model.

The test₍₀₎ positive rate was 4.0%, and the colonoscopy completion rate was 97.1% in 11594 patients who showed a positive test₍₀₎. The colonoscopy detection rate was 77.7% in those 11254 patients who underwent a colonoscopy. A total of 8748 colorectal lesions were detected (including 2182 low-risk-polyps, 2400 high-risk-polyp, and 502 colorectal cancer). The colonoscopy detection rate varied significantly with Fhb_(-1) [0 ngHb/mL: 75.6%, (0-50 ngHb/mL): 77.3%, (50-100 ngHb/mL): 88.7%, (100-150 ngHb/mL): 90.3%; P = 0.001]. People with a Fhb_(-1) within (100-150 ngHb/mL) (P = 0.001) were 2.6 (2.2; 3.0) times more likely to have a high severity level compared to those having a Fhb_(-1) value of zero. This risk was reduced by 20% in patients aged 55-59 compared to those aged < 55 [adjusted odds ratio: 0.8 (0.6; 1.0)].

The study showed that higher Fhb_(-1) is correlated to an increased risk of severity of colorectal lesions. This risk of severity increased among first-time participants (age < 55) and the elderly (≥ 70). To avoid the loss of chance in these age groups, the FIT positivity threshold should be reduced to 100 ngHb/mL. The other alternative would be to reduce the time between the two tests in these age groups from the current 2 years to 1 year.

Colorectal cancer (CRC) is the third most common cancer worldwide and the second leading cause of cancer-related death[1]. To address these major public health issues, several countries have launched nationwide CRC screening program (CRCSP) in order to reduce the incidence and mortality of CRC.

In 2015, in France, the fecal immunochemical test (FIT) OC-Sensor^® was introduced as part of the CRCSP based on its performance to detect advanced adenoma and CRC[2-5]. The positivity threshold has been set at 150 ngHb/mL (30 µgHb/g) of stool. Anyone who tested negative was re-invited for testing 2 years later. Anyone who tested positive was advised to complete a colonoscopy. Several studies showed a reduction in the CRC incidence several years after FIT introduction in the CRCSP[6,7].

Some colonic polyps such as adenomatous and serrated polyps exhibit a malignant potential, while others do not (hyperplastic, post-inflammatory, hamartomatous)[8]. Surveillance of polyp characteristic over time should lead to a significant increase in the clinical performance of the detection of polyps with high malignant potential[9]. The variability in growth behavior supports the hypothesis that some polyps continue to grow, other are remaining stable, whereas some polyps regress over time. Some authors have reported that advanced adenoma may have the potential to grow faster than non-advanced adenoma, while most small polyps remain stable or are regress over time[10,11].

It was shown that the rate of positive test results with FIT does not decrease during the subsequent campaigns, but the positive predictive value for advanced neoplasia, especially for CRC, significantly decreased in participants during the subsequent campaign after a first negative test[12,13]. A relationship between the fecal hemoglobin concentration (Fhb) and the risk to detect a CRC or an advanced adenoma has also been shown[14-21], but the severity risk of the colorectal lesions according to the earlier Fhb was poorly described.

Looking for less invasive methods for monitoring patients at moderate risk for CRC, this study aims to predict the severity of colorectal lesions based on the Fhb measured during a previous colorectal cancer screening test.

A total of 575587 patients completed at less one FIT between 2015 and 2019. Among them, 281837 (49.0%) people were excluded from this etiological analysis because they had completed only one FIT during the study duration. Finally, our study focused on the 293750 people who completed at least two FIT such as test_(-1) and test_(0). With a mean age of 61.1 ± 6.7 when completing the test_(-1), the positivity rate of the test₍₀₎ was 4.0% in these 293750 people. This positivity rate was significantly higher (P < 0.001) among patients who have a Fhb_(-1) between 100 and 150 ngHb/mL (Table 1).

A total of 282156 people (i.e. 96.2% of the sample) had a negative test₍₀₎ [Fhb₍₀₎ < 150 ngHb/mL]. The colonoscopy completion rate was globally estimated at 97.1% in 1594 people who had a positive test₍₀₎ [Fhb₍₀₎ ≥ 150 ngHb/mL]. The colonoscopy detection rate was 77.7% among those 11254 people who had a colonoscopy, with a total of 8745 colonic lesions detected (including 2182 LRP, 2400 HRP, 502 CRC). This colonoscopy detection rate varied significantly (P < 10^-3) with the Fhb_(-1) [0: 75.6%, (0-50): 77.3%, (50-100): 88.7%, (100-150): 90.3%; P < 0.001) (Table 1).

In total, 27.4% of the lesions detected (i.e. 2395 cases) were detected in people who completed a test₍₀₎ more than 30 mo after the test_(-1). However, these 2395 lesions were mostly polyps compared to the lesions detected in people who completed a test₍₀₎ within a reasonable delay between test₍₀₎ and test_(-1) (delay ≤ 24 m -LRP: 63.8%, HRP: 30.3%, CRC: 5.9%-; delay in 24-30 m -LRP: 64.2%, HRP: 29.3%, CRC: 6.5%-; delay > 30 m -LRP: 74.1%, HRP: 21.9%, CRC: 4.0%-; P = 0.0001).

For the same value of the previous Fhb [Fhb_(-1)], the colonoscopy detection rate was not significantly different between the positive values of Fhb₍₀₎ [i.e. Fhb_(-1) = 0 & Fhb₍₀₎ = (150-300): rate = 75.5% vs Fhb_(-1) = 0 & Fhb₍₀₎ ≥ 300: rate = 75.7%; P = 0.8). However, the colonoscopy detection rate was significantly higher when the Fhb₍₀₎ was higher, from 75.6% when Fhb₍₀₎ was 0 ngHb/mL to 90.3% at (100-150) ngHb/mL (Table 2).

Regardless of the age group, the colonoscopy detection rate was significantly different between the values of Fhb_(-1) [i.e. age ≥ 70 years: Fhb_(-1) = 0: rate = 79.1%, Fhb_(-1) = (0-50): Rate = 75.9%, Fhb_(-1) = (50-100): Rate = 88.2%, Fhb_(-1) = (100-150): Rate = 93.8%; P = 0.001). In the age groups of 65-69 years and ≥ 70 years, the proportion of CRC among colorectal lesions was significantly higher when the Fhb_(-1) was between 100 and 150 [i.e. age in 65-69 years: Fhb_(-1) = 0: %CRC = 6.2, Fhb_(-1) = (0-50): %CRC = 8.2, Fhb_(-1) = (50-100): %CRC = 8.1, Fhb_(-1) = (100-150): %CRC = 15.8; P = 0.001) (Table 3).

Overall, the proportion of recto-sigmoid lesions was 57.2% among the 5100 lesions for which the location was provided. The proportion of CRC among these colorectal lesions was significantly higher (P = 0.01) in cecal localization compared to other colonic locations. However, this increase in the proportion of CRC among the cecal lesions was not statistically significant regardless of the stratum combining Fhb_(-1) and Fhb₍₀₎ (Table 4).

Regardless of the Fhb_(-1) modalities [except for modality (0-50)], the variation in Fhb between test₍₀₎ and test_(-1) was significantly greater when the lesion was CRC colorectal cancer. For this modality (0-50), the number of lesions was relatively lower (CRC = 9, LRP = 27, UP = 56, HRP = 41) (Figure 1).

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Type of colorectal lesion according to fecal hemoglobin concentration variation (ng/mL) between test_(-1) and test₍₀₎ and fecal hemoglobin concentration measured at test_(-1). Fhb: Fecal hemoglobin; LRP: Low risk polyps; HRP: High risk polyps.

The multivariate logistic regression included 11254 people who completed a colonoscopy. This analysis shows that people with a Fhb_(-1) in (100-150) were 2.9 times more likely to be diagnosed with a colorectal lesion in test₍₀₎ compared to those having a Fhb_(-1) value of zero (adjusted OR = 2.9; 95%CI: 2.2-3.8). This risk of detecting a neoplastic lesion increased significantly with the age, the gender, the delay between test_(-1) and test₍₀₎, and the number of tests performed before test_(-1) (Table 5).

The ordered polytomous regression included 8748 people having exhaustive information on the colorectal lesion. The analysis showed that people with a Fhb_(-1) in (100-150) (P = 0.001) were 2.6 (2.2; 3.0) times more likely to have a high severity level, compared to those having a Fhb_(-1) value of zero. The risk to have a high severity level was higher in male compared to female (P = 0.001). This risk was reduced by 20% in age group 55-59 years compared to age group < 55 years [adjusted OR: 0.8 (0.6; 1.0), P = 0.02] (Table 5).

This study showed that the risk of detecting a colorectal lesion during a campaign was proportional to the Fhb observed in the previous test completed. Above all, it highlighted that an increase in the probability of detecting a colorectal neoplasia with a high level of severity was proportional to the Fhb observed in the previous test completed. These probabilities varied with the socio-demographic characteristics, especially with age.

The proportion of positive tests and the colonoscopy detection rate increased proportionally with the Fhb_(-1). Furthermore, the detection of colorectal lesions according to the variation in the Fhb between test_(-1) and test₍₀₎ allows to support an association between the detection of cancerous lesions and the strong Fhb variations between two consecutive tests. These results agree with previous data already described in the literature. Furthermore, the increased risk of detecting a neoplasia during a new screening 2 years after a negative screening test result has already been reported in Ireland[14] and recently in the Ile-de-France region[27]. These authors observed variable proportions of pathologies (advanced adenomas or CRC) during a subsequent campaign and hypothesized that these lesions could have been diagnosed during the previous campaign if the FIT positivity threshold was not relatively high. However, they admitted that lowering the positivity threshold would both allow the diagnosis of lesions that could be serious only 2 years later and create an unsustainable endoscopy referral burden[14].

The current positivity threshold in the French screening program induces a considerable loss of chance. Indeed, in most countries, the threshold of FIT positivity is chosen in part to adapt to the offer of colonoscopies. The need to readjust this strategy in patients with a Fhb between 100 and 150 ngHb/mL should be assessed[27]. This study highlights the need for a strategy taking into account the age of people participating in the screening campaign. The elderly (≥ 70 years) had a high proportion of positive colonoscopy when the Fhb measured in the previous campaign was between 100 and 150 ngHb/mL. Younger people (< 54 years) had a likelihood of having high-severity neoplasia. Given these results, the recommendation to reduce the FIT positivity threshold to 100 ngHb/mL for first-time arrivals and the elderly (≥ 70 years) should attract the attention of the French health authority.

Another alternative would be to reduce the delay between two tests for these first-time participants and the elderly (age ≥ 70 years) from the current 2 years to 1 year. This alternative could have its main justification in the enthusiasm of the elderly towards screening campaigns in France, described in a previous study[28]. The reduction in the time between two tests has the advantage of allowing the recovery of false negative results that a decrease in the positivity threshold cannot recover. Indeed, after finding that the Fhb was less than 4 µgHb/g in 94.0% of false negative individuals, Ibañez-Sanz et al[29] concluded that the decrease in the positivity threshold of the FIT does not increase the detection rate of advanced neoplasia but may increase costs and potential adverse effects. Whatever strategy chosen, it should also include the 65-69-age group, which accounts for more than 33.3% of the CRCs detected and presents a significantly increased proportion of CRC when Fhb_(-1) was between 100 and 150.

In terms of the location of colorectal lesions, it has been argued that FITs are possibly less effective at detecting lesions located in the proximal colon than distally[30]. Digby et al[19] showed that 77.8% of adenomas and 69.2% of cancers were located in the distal colon. The results of this study are consistent with this proximal location. However, in the Ibañez-Sanz study[29], about 60% of the lesions were localized in the proximal colon, while the expected percentage was 30%.

The optimal interval for CRC screening using FIT remains unclear[5]. In terms of the impact of the delay between the tests on the risk of detection of a colorectal lesion, the results of this study are not consistent with Van Roon's analysis, certainly because of the lower test positivity threshold (≥ 50 ng/mL) in their study carried out in a small sample (7501 people)[31]. In addition, this study found an unexpected reverse direction in the analysis of the lesion severity risk according to the delay between test₍₀₎ and test_(-1). This paradoxical reduction in the risk of severity could be explained by the high proportion of polyps among the lesions detected in patients with an abnormally long delay between test₍₀₎ and test_(-1). We can also hypothesize that some patients with a time between the two tests greater than 2 years may be symptomatic and therefore will not appear in the screening program as mentioned by Liao et al[21].

An increased risk of severity of the colorectal lesion was observed in proportion to the increase in Fhb_(-1). This risk of severity varied with the socio-demographic characteristics of the patients, especially among first-time participants. An increased colonoscopy detection rate was observed in the elderly in correlation with the increase in Fhb_(-1). According to these results, the FIT positivity threshold should be reduced to 100 ngHb/mL for first-time participants and patients aged ≥ 70. The other alternative should be to reduce the delay between the two tests for these first-time participants and the elderly (age ≥ 70) from the current 2 years to 1 year.

The authors are grateful to their colleagues from the colorectal cancer screening program management structures for the good collaboration and the quality of the data used in this study. The authors are grateful to Charlène Vandenbroucke for the thorough editing of the final version of the manuscript.