Gravitational-wave observations of binary black hole (BH) mergers provide a novel avenue for testing massive-star evolution and the resulting BH mass spectrum. Recent population analyses under the hierarchical-merger hypothesis have provided evidence for the BH mass gap and inferred its lower edge to ∼44–68 M. Motivated by these findings, we compute low-metallicity (Z = 10−5) helium star models with MESA and systematically explore the effect of uncertainties in the 12C(α, γ)16O and 16O+16O reaction rates on the final fate of these massive stars. Varying the 12C(α, γ)16O reaction rate from −3σ to +3σ, we find that the predicted BH mass gap shifts from ∼104–184 M to ∼45–135 M. In contrast, scaling the 16O+16O reaction rate by global factors of 0.1, 1, and 10 has only a modest effect on the lower edge of the BH mass gap (less than 5 M), and shifts the upper edge by more than 10 M. Using the predictions of our models together with the literature estimates for the lower edge of the BH mass gap, we constrain the astrophysical S factor of 12C(α, γ)16O reaction at 300 keV of S300 ≃ 137.6–263.4 keV barn.