An investigation into the spatial relationship between complexity and motility within the oesophagus

Evaluation of artificial neural networks in the classification of primary oesophageal dysmotility

OBJECTIVE

Artificial neural networks (ANNs) can rapidly analyse large data sets and exploit complex mathematical relationships between variables. We investigated the feasibility of utilizing ANNs in the recognition and objective classification of primary oesophageal motor disorders, based on stationary oesophageal manometry recordings.

MATERIAL AND METHODS

One hundred swallow sequences, including 80 that were representative of various oesophageal motor disorders and 20 of normal motility, were identified from 54 patients (34 F; median age 59 years). Two different ANN techniques were trained to recognize normal and abnormal swallow sequences using mathematical features of pressure wave patterns both with (ANN(+)) and without (ANN(-)) the inclusion of standard manometric criteria. The ANNs were cross-validated and their performances were compared to the diagnoses obtained by standard visual evaluation of the manometric data.

RESULTS

Interestingly, ANN(-), rather than ANN(+), programs gave the best overall performance, correctly classifying >80% of swallow sequences (achalasia 100%, nutcracker oesophagus 100%, ineffective oesophageal motility 80%, diffuse oesophageal spasm 60%, normal motility 80%). The standard deviation of the distal oesophageal pressure and propagated pressure wave activity were the most influential variables in the ANN(-) and ANN(+) programs, respectively.

CONCLUSIONS

ANNs represent a potentially important tool that can be used to improve the classification and diagnosis of primary oesophageal motility disorders.

The classification of oesophageal 24 h pH measurements using a Kohonen self-organizing feature map

Analysis of 24 h oesophageal pH studies can be problematic with many patients asymptomatic during the investigation, despite observations of reflux. The aim of this study was to carry out a cluster analysis of ambulatory pH studies to determine any underlying patterns and classes within the data. The results of 900 24 h pH studies were investigated using the Kohonen self-organizing feature map (SOFM), a neural network that can be used to identify clusters within multidimensional data. The clinical features were presented to the network and the main classes identified. The SOFM-based analysis showed that patients clinically assessed as having symptomatic reflux during the study could be described by four major classifications. The results also showed that the probability of identifying a correlation between symptoms and reflux during an investigation varies from 0.49 to 0.78 for the classes identified. The developed network may be a useful tool in the classification of pH data. The cluster-based technique may offer an alternative to standard statistical techniques for high-dimensional gastrointestinal data and form the basis of an expert system for the automated analysis of pH data.

A pH-mucosa area unit of measure to consider morphology of the oesophagus when evaluating oesophagitis

Gastro-oesophageal pH measurements are routinely carried out to quantify and determine if levels of acid reflux are responsible for symptoms. Although considered the 'gold standard', evidence suggests that pH measurements do not correlate well with the degree of oesophagitis seen during endoscopy. In this study the current measure of pH was critically examined taking into account both the effects of changes in luminal diameter and endoscopy observations. The oesophageal lumen diameter was investigated using a barium swallow for 25 patients presenting with oesophageal disorders. For each subject the widest luminal diameter was measured for a series of five controlled swallows. The results showed that the lumen diameter varied widely from 0.9 to 3.8 cm. An alternative approach to the current measurement of pH was explored. In this approach the exposure not only included the luminal pH and time exposed but also the area of mucosa exposed as a result of differing luminal diameters. Although it is currently not possible to assess the diameter or morphology of the oesophageal lumen during a pH study, the analysis highlighted that the current measure of pH exposure time does not include the area of mucosa exposed. These results may explain, to some extent, the poor correlation between pH measurements and degree of oesophagitis seen during endoscopy.