| Characteristic | Negative, N = 45,5741 | False Positive, N = 3,3531 | p-value2 |
|---|---|---|---|
| Age at Cologuard test (years) | 58 (50, 65) | 63 (56, 69) | <0.001 |
| Sex | 0.4 | ||
| Female | 27,350 (60%) | 1,979 (59%) | |
| Male | 18,195 (40%) | 1,371 (41%) | |
| Uncertain/Unknown | 29 (<0.1%) | 3 (<0.1%) | |
| Race | <0.001 | ||
| WHITE | 40,802 (90%) | 3,169 (95%) | |
| BLACK OR AFRICAN AMERICAN | 1,791 (3.9%) | 77 (2.3%) | |
| ASIAN OR PACIFIC ISLANDER | 1,436 (3.2%) | 42 (1.3%) | |
| AMERICAN INDIAN OR ALASKA NATIVE | 277 (0.6%) | 12 (0.4%) | |
| UNKNOWN | 1,268 (2.8%) | 53 (1.6%) | |
| Ethnicity | <0.001 | ||
| Not Hispanic or Latino | 42,224 (93%) | 3,231 (96%) | |
| Hispanic/Latino | 2,182 (4.8%) | 70 (2.1%) | |
| Unknown | 1,168 (2.6%) | 52 (1.6%) | |
| BMI | 29 (25, 34) | 29 (25, 35) | 0.4 |
| Unknown | 25,543 | 1,877 | |
| Any other cancer | 669 (1.5%) | 145 (4.3%) | <0.001 |
| Unknown | 2 | 0 | |
| Head and neck cancer | 166 (0.4%) | 21 (0.6%) | 0.018 |
| Unknown | 2 | 0 | |
| Lung cancer | 227 (0.5%) | 50 (1.5%) | <0.001 |
| Unknown | 2 | 0 | |
| Non-colorectal GI cancer | 276 (0.6%) | 74 (2.2%) | <0.001 |
| Unknown | 2 | 0 | |
| 1 Median (IQR); n (%) | |||
| 2 Wilcoxon rank sum test; Fisher’s exact test; Pearson’s Chi-squared test | |||
Assessing the Risk of Non-Colorectal Malignancy in Patients with a ``False Positive’’ Cologuard Result
Statistical analysis
Categorical variables were summarized as counts and percentages, and continuous variables were summarized as medians and interquartile ranges (IQRs). Baseline characteristics were compared using the chi-square or Fisher’s exact test for categorical variables and the Wilcoxon rank-sum test for continuous variables. Kaplan–Meier plots and log-rank tests were used to compare cancer-free survival (CFS) and overall survival (OS) by Cologuard result. Cox proportional hazards regression models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) of CFS and OS by Cologuard result adjusting for demographics. A two-sided p value < 0.05 was considered statistically significant. All statistical analyses were performed using R (version 4.3.1).
Results
Patient characteristics
The cohort consists of 48927 patients, including 45574 with negative and 3353 with false positive Cologuard results.
Baseline characteristics and cancer rates are summarized in Table 1 by Cologuard result.
- Patients with false positive Cologuard results have higher incidences of non-colorectal cancers.
CFS and OS
Follow-up and crude event rates of cancer-free survival (primary endpoint) and overall survival are summarized in Table 2 by Cologuard result.
| Characteristic | Negative, N = 45,5741 | False Positive, N = 3,3531 |
|---|---|---|
| CFS Follow-up (months) | 32 (16, 51) | 34 (17, 56) |
| Cancer-free survival event | 1,469 (3.2%) | 216 (6.4%) |
| OS Follow-up (months) | 33 (16, 51) | 36 (18, 57) |
| Deceased | 898 (2.0%) | 100 (3.0%) |
| 1 Median (IQR); n (%) | ||
Kaplan–Meier plots and log-rank tests of cancer-free survival (CFS) and overall survival (OS) by Cologuard result are shown in Figure 1. Patients with false positive Cologuard results have significantly worse CFS (p < 0.001) and OS (p=0.008).
Survival estimates at 24, 48, and 72 months for CFS and OS survival by Cologuard result are summarized in Table 3.
| Endpoint | Month 24 | Month 48 | Month 72 |
|---|---|---|---|
| CFS | |||
| Negative | 98.2% (98.1%, 98.4%) | 95.7% (95.5%, 96.0%) | 92.0% (91.4%, 92.5%) |
| False Positive | 96.0% (95.3%, 96.7%) | 92.7% (91.6%, 93.9%) | 87.0% (84.9%, 89.2%) |
| OS | |||
| Negative | 99.1% (99.0%, 99.2%) | 97.3% (97.1%, 97.5%) | 94.9% (94.5%, 95.3%) |
| False Positive | 98.3% (97.8%, 98.8%) | 96.4% (95.6%, 97.2%) | 93.7% (92.2%, 95.2%) |
Cox regression analyses
Cox proportional hazards regression analyses of CFS and OS against Cologuard result adjusting for demographics are summarized in Table 4.
| Characteristic | Cancer-free survival | Overall survival | ||||
|---|---|---|---|---|---|---|
| HR1 | 95% CI1 | p-value | HR1 | 95% CI1 | p-value | |
| Cologuard result | ||||||
| Negative | — | — | — | — | ||
| False Positive | 1.52 | 1.31, 1.75 | <0.001 | 1.09 | 0.89, 1.34 | 0.4 |
| Age at Cologuard test (years) | 1.06 | 1.06, 1.07 | <0.001 | 1.08 | 1.07, 1.09 | <0.001 |
| Sex | ||||||
| Female | — | — | — | — | ||
| Male | 1.45 | 1.31, 1.59 | <0.001 | 1.54 | 1.36, 1.75 | <0.001 |
| Uncertain/Unknown | 1.33 | 0.19, 9.45 | 0.8 | 0.00 | 0.00, Inf | >0.9 |
| Race | ||||||
| WHITE | — | — | — | — | ||
| BLACK OR AFRICAN AMERICAN | 1.40 | 1.08, 1.82 | 0.010 | 1.49 | 1.07, 2.08 | 0.020 |
| ASIAN OR PACIFIC ISLANDER | 1.02 | 0.73, 1.42 | >0.9 | 0.71 | 0.42, 1.20 | 0.2 |
| AMERICAN INDIAN OR ALASKA NATIVE | 1.30 | 0.68, 2.51 | 0.4 | 1.78 | 0.85, 3.74 | 0.13 |
| UNKNOWN | 0.65 | 0.43, 0.97 | 0.037 | 0.70 | 0.42, 1.16 | 0.2 |
| 1 HR = Hazard Ratio, CI = Confidence Interval | ||||||
(Ethnicity is non-significant after adjusting for race.)
- After adjusting for demographics:
- False positives have a significantly higher risk of non-colorectal cancer or death (CFS: HR 1.52, 95% CI 1.31–1.75, p < 0.001).
- False positives also have a higher risk of death (p < 0.4)
- Statistical nonsignificance may be due to adjustment of age (FP are older; see Table 1).
- Factors significantly associated with worse CFS and/or OS:
- Older age
- Male vs female
- Black or African American vs White