Sebastian Foersch

Supplementary Table 2 shows the pooled-arm analyses isolating the effects of aspirin and metformin on global NFκB expression, adjusted for baseline values.

Supplementary Table 1 shows baseline immunohistochemical expression of NFκB, EGFR, Ki67 and pS6K in normal right and left colonic mucosa by treatment arm.

Main baseline patient characteristics.

Supplementary Table 4 describes the representativeness of the study participants with respect to sex, age, race/ethnicity and geography.

Number of adverse events and grade by treatment arm.

Participant flow diagram including surrogate endpoints.

Supplementary Table 3 shows the pooled-arm analyses isolating the effects of aspirin and metformin on global Ki67 expression, adjusted for baseline values.

Purpose: Aspirin (Asa) and metformin (Met) have shown preventive efficacy in colorectal cancer. Preclinical data suggest convergence on NF-κB/PI3K–mTOR signaling and potential synergy. The ASAMET phase II trial tested this hypothesis in patients with resected stage I to III colorectal cancer. Patients and Methods: In a 2 × 2 factorial, multicenter, double-blind study (EudraCT 2015-004824-77; NCT03047837), 51 patients were randomized to placebo, Asa 100 mg once daily, Met 850 mg twice daily, or AsaMet for 12 months. Biopsies of normal colonic mucosa and fasting blood samples were obtained at baseline and 12 months. NF-κB (primary endpoint), Ki67, EGFR, and pS6K were assessed by IHC. Results: After 12 months, compared with placebo, Met reduced global NF-κB by −9.04 [95% confidence interval (CI), −17.55 to −0.54; P = 0.04], whereas Asa alone was inactive (−3.58; 95% CI, −12.15 to 4.99; P = 0.41), and AsaMet showed a nonsignificant reduction (−8.88; 95% CI, −18.86 to 1.10; P = 0.08). Pooled analyses confirmed a Met-driven effect (Met vs. non-Met: −7.29; 95% CI, −13.79 to −0.79; P = 0.03), with no impact of Asa (P = 0.73). Ki67 and pS6K were unchanged globally; however, AsaMet increased right-sided Ki67 (12.48; 95% CI, 0.45–24.51; P = 0.04). No treatment modulated pS6K. Toxicity was manageable, and no serious adverse events occurred. Conclusions: Met reduces NF-κB expression in normal colorectal mucosa of survivors of colorectal cancer, supporting its potential as a low-toxicity preventive agent. Asa did not favorably affect mucosal biomarkers. These findings support tissue-based endpoints and further evaluation of Met for colorectal cancer prevention. Significance: The ASAMET trial provides the first randomized evidence that Met significantly modulates NF-κB expression in histologically normal colorectal mucosa of survivors of colorectal cancer. This finding supports a biologically plausible mechanism of action in secondary cancer prevention that could be exploited in the postadjuvant setting. Asa, by contrast, did not favorably affect any mucosal biomarker endpoint in this population. These results may support a phase III study of Met in this setting. Trial Registration: ClinicalTrials.gov Identifier: NCT03047837; EudraCT number: 2015-004824-77.

NonMSIH to MSIH counterfactuals with cell segmentation