Morphologic Characterization of Pancreatic Ductal Adenocarcinoma following Post-neoadjuvant Pancreatectomy and Clinical Value of Intratumor Heterogeneity

OBJECTIVE: To evaluate the morphologic landscape of pancreatic ductal adenocarcinoma (PDAC), intratumor spatial heterogeneity, and the resulting clinical impact following post-neoadjuvant pancreatectomy. BACKGROUND: The clinical value of PDAC morphologic subtypes and intratumor spatial heterogeneity posttreatment remains an open issue. METHODS: The study cohort included patients who underwent post-neoadjuvant pancreatectomy for PDAC at the University of Verona Hospital Trust between 2013 and 2019. All hematoxylin and eosin-stained slides were reviewed to assess PDAC histomorphology and intratumor heterogeneity. The relationship with other clinicopathological variables, overall survival, and recurrence-free survival was evaluated using standard statistics. RESULTS: The study cohort included 400 patients. Histologic revision identified 10 different morphologic subtypes. Gland-forming PDAC with a conventional pattern was the most frequently identified subtype (41.8%). Overall, 247 tumors (61.7%) showed only one histologic pattern and were classified as homogeneous, whereas 153 (38.3%) showed different morphologies and were classified as heterogeneous tumors. The median postresection survival was 30.1 months (95% CI: 26.6-33.5). There was a substantial survival variability according to the morphologic subtype, ranging from 19.1 months in the gyriform subtype to 47.0 months in the papillary subtype. Tumors with a heterogeneous morphology displayed a higher rate of nodal metastases, worse tumor regression metrics, and worse oncologic outcomes relative to spatially homogeneous tumors. CONCLUSIONS: This paper provided a morphologic taxonomy of residual tumors following post-neoadjuvant pancreatectomy for PDAC. The morphologic subtype and intratumor spatial heterogeneity have relevant prognostic implications and could be included in the pathology report to complement regression metrics.