Noninvasive feature-based detection of delayed gastric emptying in humans using neural networks

Radioscintigraphy is currently the gold standard for gastric emptying test which involves radiation exposure and is considerably expensive, We present a feature-based detection approach using neural networks for the noninvasi ve diagnosis of delayed gastric emptying from the cutaneous electrogastrogr am (EGG). Simultaneous recordings of the EGG and scintigraphic gastric empt ying test were made in 152 patients with symptoms suggestive of delayed gas tric emptying. Spectral analyses were performed to derive EGG parameters wh ich were used as the input of the neural network. The result of scintigraph ic gastric emptying was used as the gold standard for the training and test ing of the neural network. A correct classification of 85% (a specificity o f 89% and a sensitivity of 82%) was achieved using the proposed method.