Contrast-enhanced harmonic endoscopic ultrasound (CH-EUS) has shown high sensitivity and specificity in the diagnosis of solid pancreatic lesions. However, its visual interpretation is subjective and lacks a standardized method. Therefore, quantitative analyses such as time–intensity curves (TIC) have been proposed to provide a more objective assessment of lesions. The primary objective was to evaluate whether TIC analysis using CH-EUS can characterize pancreatic tumors. The secondary objective was to determine the most accurate parameter for this purpose.

This retrospective study included 67 patients with pancreatic tumors evaluated by CH-EUS between 2019 and 2025. TIC were generated to show changes in signal intensity over time, evaluating four parameters: peak intensity, time to peak intensity, intensity at 60 seconds after peak, and washout rate. Histological or cytological diagnosis was considered the “gold standard” and used to divide patients into three groups: pancreatic carcinomas (n=54), pancreatic neuroendocrine tumors (PNETs) (n=10), and other lesions (n=3).

During the study period, 67 patients with solid pancreatic lesions were included, 28 men and 39 women, with a mean age of 70 years.There were 54 pancreatic adenocarcinomas, 10 neuroendocrine tumors, and among the 3 patients in the “other tumors” group, there was a plasma cell neoplasm with kappa light-chain restriction, metastasis from a primary breast tumor, and metastasis from a renal clear cell tumor.

Significant differences were found in the qualitative analysis of contrast uptake, with iso- or hyperenhancement associated with neuroendocrine tumors, and hypoenhancement associated with pancreatic adenocarcinoma.

Regarding quantitative analysis, no statistically significant differences were found in time to peak intensity or washout rate between adenocarcinomas and neuroendocrine tumors.Peak intensity and intensity at 60 seconds were significantly higher in PNETs than in adenocarcinomas (p < 0.05).

The areas under the ROC curve (AUC) for diagnosing PNETs using peak intensity and intensity at 60 seconds were 0.692 and 0.691, respectively, with cutoff values of 61.45 dB and 16.895 dB.

Statistically significant differences were also found regarding lesion size, obtaining a cutoff point of 22.5 mm after performing the ROC curve analysis. Therefore, when lesion size was included in the model, the areas under the ROC curve for diagnosing NET using peak intensity and intensity at 60 seconds were 0.994 and 0.986, respectively.

Peak intensity was the most accurate TIC parameter for differentiating PNETs, with a sensitivity of 98.3%% and specificity of 100%.

TIC analysis using CH-EUS demonstrated marked differences between pancreatic carcinomas and neuroendocrine tumors. Peak intensity is the most accurate diagnostic parameter for differentiating pancreatic tumors. In resectable lesions without the need for biopsy, confirmation of our results in prospective, blinded, multicenter studies could expand the preoperative understanding of these lesions.