This study aimed to explore the effects of different breath-holding times (20, 18 and 15 s) on PET/computed tomography (PET/CT) image quality and metabolic parameters of hypermetabolic lung nodules. We hope that our findings will elucidate the clinical feasibility of short-duration breath-hold PET imaging. Compared to free-breathing PET group, all breath-hold groups showed significantly higher tumour-to-background ratio but lower signal-to-noise ratio (all P < 0.05). Importantly, the 18 s breath-hold group demonstrated overall image quality, fusion scores and contrast-to-noise ratio comparable to those of the 20-s group, without statistically significant differences (all P > 0.05). Moreover, the 18-s group exhibited a more favourable signal-to-noise ratio profile than the 15-s group. Metabolically, all breath-hold scans significantly increased maximum standardized uptake value (SUVmax) and mean SUV while reducing metabolic tumour volume and total lesion glycolysis relative to free-breathing (P < 0.05). Notably, there were no significant differences in key quantitative metrics, including maximum SUV, peak SUV, metabolic tumour volume and total lesion glycolysis, among the three breath-hold groups. The consistency of peak SUV across groups further supports its reliability as a robust parameter under reduced acquisition times. The 18-s breath-hold protocol is a valid clinical alternative to the 20-s acquisition, offering comparable image quality and metabolic data while superior efficiency and patient tolerability, making it an ideal choice for routine practice.