The Ki67 index evaluation of pancreatic neuroendocrine tumors using 3D immunohistochemistry in small tissue specimens

/Objectives: This study aimed to evaluate the usefulness of three-dimensional (3D) immunohistochemistry for the Ki67 index of small tissue specimens of pancreatic neuroendocrine tumor (PanNET). Clinicopathological materials from 17 patients with PanNET who underwent surgical resection at Jichi Medical University Hospital were analyzed. We compared the Ki67 index of endoscopic ultrasonography-fine-needle aspiration biopsy (EUS-FNAB) specimens, surgical specimens, and small tissue specimens hollowed from paraffin blocks of surgical specimens that were substituted for EUS-FNAB specimens (“sub-FNAB”). The sub-FNAB specimens were optically cleared using LUCID (IlLUmination of Cleared organs to IDentify target molecules) and analyzed using 3D immunohistochemistry. The median Ki67 index in FNAB, sub-FNAB, and surgical specimens with conventional immunohistochemistry were 1.2% (0.7–5.0), 2.0% (0.5–14.6), and 5.4% (1.0–19.4), respectively. The median Ki67 index in sub-FNAB specimens with tissue clearing was calculated separately using the total number of cells on multiple images (“multiple slice”), with the image of the fewest positive cells (“coldspot”), and with the image of most positive cells (“hotspot”), which were 2.7% (0.2–8.2), 0.8% (0–4.8), and 5.5% (2.3–12.4), respectively. PanNET grade evaluated for the hotspot of the surgical specimens was significantly more consistent with those of the hotspot than multiple images of sub-FNAB specimens (16/17 vs. 10/17, p = 0.015). Hotspot evaluation using 3D immunohistochemistry of the sub-FNAB specimens showed agreement with the assessment of the surgical specimens (Kappa coefficient: 0.82). Tissue clearing and 3D immunohistochemistry for the Ki67 index can potentially improve the preoperative evaluation of EUS-FNAB specimens of PanNET in routine clinical practice.