Abstract: Objective: To provide a precise quantitative basis for computed tomography perfusion (CTP) quality control by investigating the impact of arterial input function (AIF) washout phase truncation on the accuracy of CTP results, establishing criteria for determining AIF curve integrity, and constructing the first quantitative relationship model between AIF peak and post-baseline levels. Methods: A retrospective analysis was conducted on 34 patients with suspected acute anterior circulation ischemic stroke who underwent CTP from January 2023 to January 2024, and had complete AIF curves (post-baseline duration ≥18 s). The first low point of the descending branch after the AIF peak was defined as time zero. The perfusion parameters corresponding to the complete 18 s post-zero data were used as the “gold standard”. The “optimal scan duration” (defined as the latest truncation time with a relative deviation of parameters from the gold standard ≤10%) was determined using a stepwise simulated truncation method. Linear regression was applied to analyze the correlation between the AIF peak CT value and post-baseline average CT value. Results: The “optimal scan duration” in 79.4% (27/34) of patients was after time zero. This metric showed a skewed distribution. Quantile analysis revealed that the optimal continuous scanning times after time zero corresponding to the 50th percentile (P50), 75th percentile (P75, recommended clinical value), and 90th percentile (P90) were 9.0 s, 12.0s, and 15.0 s, respectively. A strong positive correlation was found between the AIF peak CT value and post-baseline average CT value (r=0.800, P < 0.001). The regression model was as follows: Post-baseline CT value=0.18×AIF peak CT value+23.59 (R²=0.639, P < 0.001). Conclusion: Complete acquisition of the AIF washout phase requires continuous scanning for at least 12 s after time zero to encompass 75% of patients for routine quality control, while 15 s can cover more than 90% of patients. The established regression model provides a quantitative tool for evaluating AIF curve integrity and further offers a theoretical basis for future intelligent AI-based quality control.