A novel histopathological classification of implant periapical lesion: A systematic review and treatment decision tree

Abstract

Background Implant periapical lesion (IPL), as a peri-implant disease originating from implant apex, maintains coronal osseointegration in the early stage. With the understanding to IPL increasingly deepened, IPL classification based on different elements was proposed although there still lacks an overall classification system. This study, aiming to systematically integrate the available data published in the literature on IPL associated with histopathology, proposed a comprehensive classification framework and treatment decision tree for IPL. Methods and findings English articles on the topic of “implant periapical lesion”, “retrograde peri-implantitis” and “apical peri-implantitis” were searched on PubMed, Embase and Web of Science from 1992 to 2021, and citation retrieval was performed for critical articles. Definite histopathology and radiology of IPL are indispensable criteria for including the article in the literature. The protocol was registered in PROSPERO (CRD42022378001). A total of 509 papers identified, 28 studies were included in this review. In only one retrospective study, 37 of 39 IPL were reported to be at the inflammatory or abscess stage. 27 cases (37 implants) were reported, including acute non-suppurative (1/37, developed to chronic granuloma), chronic granuloma (5/37), acute suppurated (2/37), chronic suppurated-fistulized (6/37), implant periapical cyst (21/37), poor bone healing (2/37), foreign body reaction (1/37). Antibiotics alone did not appear to be effective, and the consequence of surgical debridement required cautious interpretation because of the heterogeneity of lesion course and operation. Implant apicoectomy and marsupialization were predictable approaches in some cases. Conclusions The diversiform nature of IPL in the case reports confirms the need for such histopathological classification, which may enhance the comparison and management of different category.

Ensemble learning based transmission line fault classification using phasor measurement unit (PMU) data with explainable AI (XAI)

“…The classification of transmission line disturbances [ 22 , 53 ] has been performed using nine distinct supervised multiclass techniques, which have been enlisted in Table 6 . Decision Tree (DT) [ 54 ], Random Forest (RF) [ 55 ], and K-Nearest Neighbor (K-NN) [ 56 , 57 ] have achieved more than 99% train and test accuracy. Decision Tree (DT), Random Forest (RF), and K-Nearest Neighbor (K-NN) have outperformed the other classifiers in the classification of six labels of short circuit fault scenarios.…”