Artificial neural network classification of pharyngeal high‐resolution manometry with impedance data

Modifying the ANN architecture, such as increasing the number of hidden nodes, had a minimal effect on classification accuracy.

“Our efforts to improve performance by modifying the architecture of the ANN, such as increasing the number of hidden nodes, had a minimal effect on classification accuracy (table 4).”

Maximum velopharynx pressure achieved the highest individual classification accuracy at 38.6%.

“Maximum velopharynx pressure achieved the highest classification accuracy at 38.6%.”

• Also stated in: § DISCUSSION · plain

  “Maximum velopharyngeal pressure (38.6%), velopharyngeal pressure integral (36.3%), and post-closure UES rise time (33.1%) provided the highest classification accuracies.”

No single parameter was able to classify swallows as penetration or aspiration with high accuracy.

“More interestingly, no single parameter was able to classify swallows as penetration or aspiration with high accuracy.”

For aspiration, combining all parameters yielded a classification accuracy of 82.5%.

“combining all parameters led to significantly higher classification rates of 65.9% for penetration and 82.5% for aspiration.”

ROC analysis yielded an area under the curve of 0.8912 for safe swallows.

“Receiver operating characteristic (ROC) analysis yielded areas under the curve (AUC) of 0.8912 for safe, 0.8187 for aspiration, and 0.8014 for penetration.”

• Also stated in: § Artificial neural network classification · plain

  “ROC analysis yielded curves with AUC of 0.8912 for safe (figure 2a), 0.8187 for aspiration (figure 2b), and 0.8014 for penetration (figure 2c).”

The majority of unsafe swallows were on a liquid bolus (38 of 49).

“As the majority of unsafe swallows were on a liquid bolus (thirty-eight of forty-nine)”

Bolus characteristics did not appear to negatively affect classification accuracy in this experiment.

“Bolus characteristics did not appear to negatively affect classification accuracy in this experiment”

• AssumptionRests on an assumption: A main benefit of ANN analysis is that it recognizes patterns rather than trends in single variables.

When individual parameters were used, overall classification accuracy ranged from 63.8% to 73.6%.

“Percent accuracy ranged from 63.8% to 73.6% (table 3).”

• Also stated in: § DISCUSSION · plain

  “Percent classification accuracy ranged from 79.1% to 98.2% for safe swallows and from 63.8% to 73.6% overall.”

The artificial neural network classified swallows with an accuracy of 89.4±2.4% when all parameters were included.

“A classification accuracy of 89.4±2.4% was achieved when including all parameters.”

• Also stated in: § Artificial neural network classification · plain

  “A classification accuracy of 89.4±2.4% was achieved when including all parameters.”

Individual parameters had low classification rates for penetration and aspiration, ranging from 7.1% to 38.6%.

“Individual parameters had relatively low classification rates for penetration and aspiration, with a range of 7.1% to 38.6% (table 3).”

• Also stated in: § DISCUSSION · plain

  “Percent classification accuracy ranged from 7.1% to 38.6%.”

The artificial neural network classified swallows with an accuracy of 76.0±4.9% when only impedance-related parameters were included.

“Including only manometry-related parameters yielded a classification accuracy of 85.0±6.0% while including only impedance-related parameters yielded a classification accuracy of 76.0±4.9%.”

• Also stated in: § Artificial neural network classification · plain

  “Including only manometry-related parameters yielded a classification accuracy of 85.0±6.0% while including only impedance-related parameters yielded a classification accuracy of 76.0±4.9% (table 1).”

The artificial neural network classified swallows with an accuracy of 85.0±6.0% when only manometry-related parameters were included.

“Including only manometry-related parameters yielded a classification accuracy of 85.0±6.0% while including only impedance-related parameters yielded a classification accuracy of 76.0±4.9%.”

• Also stated in: § Artificial neural network classification · plain

  “Including only manometry-related parameters yielded a classification accuracy of 85.0±6.0% while including only impedance-related parameters yielded a classification accuracy of 76.0±4.9% (table 1).”

ROC analysis yielded an area under the curve of 0.8014 for penetration.

“Receiver operating characteristic (ROC) analysis yielded areas under the curve (AUC) of 0.8912 for safe, 0.8187 for aspiration, and 0.8014 for penetration.”

• Also stated in: § Artificial neural network classification · plain

  “ROC analysis yielded curves with AUC of 0.8912 for safe (figure 2a), 0.8187 for aspiration (figure 2b), and 0.8014 for penetration (figure 2c).”

The three impedance parameters (PNadImp, TPNadImp-PeakP, pharyngeal flow interval) differed significantly or near-significantly across the safe, penetration, and aspiration groups.

“All impedance parameters were significantly or near significantly different across groups (PNadImp: p=0.015; TPNadImp-PeakP: p=0.054; pharyngeal flow interval: p<0.001).”

• Also stated in: § DISCUSSION · plain

  “Each of the three impedance parameters (PNadImp, TPNadImp-PeakP, pharyngeal flow interval) was significantly different across categories.”

There was not a consistent correlation between ANOVA p-values and classification accuracy.

“Most notably, there was not a consistent correlation between p-values resulting from ANOVA and classification accuracy.”

ROC analysis yielded an area under the curve of 0.8187 for aspiration.

“Receiver operating characteristic (ROC) analysis yielded areas under the curve (AUC) of 0.8912 for safe, 0.8187 for aspiration, and 0.8014 for penetration.”

• Also stated in: § Artificial neural network classification · plain

  “ROC analysis yielded curves with AUC of 0.8912 for safe (figure 2a), 0.8187 for aspiration (figure 2b), and 0.8014 for penetration (figure 2c).”

Many of the significant manometric parameters were related to the UES.

“Of the manometric parameters demonstrating significant differences, many were related to the UES.”

Classification accuracy was higher for aspiration than for penetration.

“Classification accuracy was higher for aspiration than penetration.”

The low number of unsafe swallows (17 penetration, 32 aspiration) may have contributed to low classification rates.

“Low classification rates could be due to difficulty distinguishing between the two types of unsafe swallows and could also be due to the relatively low number of unsafe swallows (17 penetration, 32 aspiration) included in the analysis.”

Differences detected in individual parameters across groups cannot be completely attributed to swallow status.

“differences detected in individual parameters across groups cannot be completely attributed to swallow status (i.e., safe, penetration, or aspiration).”

Neither manometric nor impedance parameters alone provided penetration classification accuracy high enough to provide a basis for clinical decision-making.

“Neither rate is high enough to provide a basis for clinical decision-making”

Including only manometric parameters led to greater classification accuracy than including only impedance parameters.

“Interestingly, including only manometric parameters led to greater classification accuracy than including only impedance parameters.”

The greatest classification accuracy was achieved when combining both pressure and impedance information.

“The greatest classification accuracy, though, was achieved when combining both pressure and impedance information, demonstrating the importance of a comprehensive assessment.”

The disparity between individual parameter classification rates and group classification rates demonstrates the importance of multi-parameter assessment.

“This disparity between individual parameter classification rates and group classification rates demonstrates the importance of multi-parameter assessment.”

Manometric parameters describing UES muscular activity are strong predictors of pharyngeal function.

“Manometric parameters describing the muscular activity in this region are strong predictors of pharyngeal function, as an inability to pass the bolus through the UES due to poorly controlled muscular activity can result in penetration and aspiration of materials retained in the pharynx.26-27”

• cites [26] sourceNot checked yet

  Not checked in the free run.

• cites [27] sourceNot checked yet

  Not checked in the free run.

The pattern recognition techniques employed here appear to be clinically useful in identifying patients at risk for aspiration.

“The pattern recognition techniques employed here appear to be clinically useful in identifying patients at risk for aspiration.”

• AssumptionRests on an assumption: Detecting aspiration is critical to the clinical utility of any swallowing evaluation.

Individual parameter analyses are more susceptible to identifying differences due to bolus type rather than swallow type.

“Individual parameter analyses are more susceptible to identifying differences due to bolus type rather than swallow type.”

The high accuracy in this preliminary study provides evidence that HRM-impedance has potential as an alternative clinical assessment tool, especially when coupled with ANN techniques.

“The high accuracy in this preliminary study provides evidence that HRM-impedance has potential as an alternative clinical assessment tool, especially when coupled with ANN techniques.”

• AssumptionRests on an assumption: HRM-impedance provides an alternative assessment that avoids radiation, can be done at the bedside, and yields quantitative information on many swallowing-related parameters.

Classification models show high accuracy in classifying swallows from dysphagic patients as safe or unsafe.

“Classification models show high accuracy in classifying swallows from dysphagic patients as safe or unsafe.”

Manometric and impedance parameters performed better collectively than either group did individually.

“Manometric and impedance parameters performed better collectively than either group did individually, as demonstrated by the differences in classification rates (table 1).”

Combining pressure measurement with impedance measurement provides an assessment of both motility and flow of luminal content.

“Combining pressure measurement with impedance measurement provides an assessment of both motility and flow of luminal content.5”

• cites [5] sourceNot checked yet

  Not checked in the free run.

HRM-impedance with ANN represents one method which could be used clinically to screen for patients at risk for penetration or aspiration.

“HRM-impedance with ANN represents one method which could be used clinically to screen for patients at risk for penetration or aspiration.”

• AssumptionRests on an assumption: HRM-impedance provides an alternative assessment that avoids radiation, can be done at the bedside, and yields quantitative information on many swallowing-related parameters.
• AssumptionRests on an assumption: Detecting aspiration is critical to the clinical utility of any swallowing evaluation.
• AssumptionRests on an assumption: Penetration and aspiration are common in dysphagic patients because of weakness and poorly coordinated swallowing.

This provides support for a comprehensive multi-parameter analysis such as that provided by ANN analysis of HRM-impedance, rather than isolated measurements of peak pharyngeal pressure or UES pressure.

“Accordingly, this provides support for a comprehensive multi-parameter analysis such as that provided by ANN analysis of HRM-impedance, rather than isolated measurements of peak pharyngeal pressure or UES pressure, for example.”

An inability to pass the bolus through the UES due to poorly controlled muscular activity can result in penetration and aspiration of materials retained in the pharynx.

“Manometric parameters describing the muscular activity in this region are strong predictors of pharyngeal function, as an inability to pass the bolus through the UES due to poorly controlled muscular activity can result in penetration and aspiration of materials retained in the pharynx.26-27”

• cites [26] sourceNot checked yet

  Not checked in the free run.

• cites [27] sourceNot checked yet

  Not checked in the free run.

The low number of penetration swallows (seventeen) may have contributed to the low penetration classification rate.

“The low number of swallows exhibiting penetration included in the analysis (seventeen) may also have contributed to the low classification rate, as introducing fewer examples of one category during the training stage can have an adverse effect during the testing stage.”

The relationships among variables, rather than absolute values of each parameter in isolation, appear to be the key in distinguishing safe from unsafe swallows.

“This finding also has physiological significance, as the relationships among the variables of interest, rather than absolute values of each parameter in isolation, appear to be the key in distinguishing safe from unsafe swallows.”