Cardiotocography (CTG) consists in the simultaneous recording of two distinct traces, the fetal heart rate (FHR; bpm) and the maternal uterine contractions (UCs; mmHg). CTG analysis consists in the evaluation of specific features of traces, among which fetal decelerations (DECs) are considered the 'center-stage' since possibly related to fetal distress. DECs are classified based on their duration and occurrence in relation to UCs as prolonged, early, late and variable; each class associates to a specific status of the fetus health. Typically, CTG traces are visually interpreted; however, computerized CTG analysis may overcome subjectivity in CTG interpretation. Thus, this study proposes a new automatic algorithm for computerized identification and classification of DECs. The algorithm was tested on the 552 CTG recordings constituting the 'CTU-CHB intra-partum CTG database' of Physionet. Of these, 470 (85.15%) were found suitable for automatic DECs identification and classification. Overall, 5888 DECs were identified, of which 3255 (55.28%) were classified while the other 2633 (44.72%) remained unclassified due to very strict preliminary classification criteria (now required for avoiding misclassifications). Among the classified DECs, 468 (14.38%) were classified as prolonged, 1498 (46.02%) as early, 32 (0.98%) as late, 1257 (38.62%) as variable. Thus, among the classified DECs, the most common are the early and the variable ones (overall 84.64%), the occurrence of which ranged from 0 to 14 DECs per recording. These findings are in agreement with what reported in literature. In conclusion, the proposed algorithm for automatic DECs identification and classification represents a useful tool for computerized CTG analysis.
Automatic Identification and Classification of Fetal Heart-Rate Decelerations from Cardiotocographic Recordings / Sbrollini, Agnese; Carnicelli, Amalia; Massacci, Alessandra; Tomaiuolo, Leonardo; Zara, Tommaso; Marcantoni, Ilaria; Burattini, Luca; Morettini, Micaela; Fioretti, Sandro; Burattini, Laura. - ELETTRONICO. - 2018-:(2018), pp. 474-477. (Intervento presentato al convegno 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2018 tenutosi a Hawaii Convention Center, usa nel 2018) [10.1109/EMBC.2018.8512432].
Automatic Identification and Classification of Fetal Heart-Rate Decelerations from Cardiotocographic Recordings
Sbrollini, Agnese;Marcantoni, Ilaria;Morettini, Micaela;Fioretti, Sandro;Burattini, Laura
2018-01-01
Abstract
Cardiotocography (CTG) consists in the simultaneous recording of two distinct traces, the fetal heart rate (FHR; bpm) and the maternal uterine contractions (UCs; mmHg). CTG analysis consists in the evaluation of specific features of traces, among which fetal decelerations (DECs) are considered the 'center-stage' since possibly related to fetal distress. DECs are classified based on their duration and occurrence in relation to UCs as prolonged, early, late and variable; each class associates to a specific status of the fetus health. Typically, CTG traces are visually interpreted; however, computerized CTG analysis may overcome subjectivity in CTG interpretation. Thus, this study proposes a new automatic algorithm for computerized identification and classification of DECs. The algorithm was tested on the 552 CTG recordings constituting the 'CTU-CHB intra-partum CTG database' of Physionet. Of these, 470 (85.15%) were found suitable for automatic DECs identification and classification. Overall, 5888 DECs were identified, of which 3255 (55.28%) were classified while the other 2633 (44.72%) remained unclassified due to very strict preliminary classification criteria (now required for avoiding misclassifications). Among the classified DECs, 468 (14.38%) were classified as prolonged, 1498 (46.02%) as early, 32 (0.98%) as late, 1257 (38.62%) as variable. Thus, among the classified DECs, the most common are the early and the variable ones (overall 84.64%), the occurrence of which ranged from 0 to 14 DECs per recording. These findings are in agreement with what reported in literature. In conclusion, the proposed algorithm for automatic DECs identification and classification represents a useful tool for computerized CTG analysis.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.