Limits...
Automatic Detection of Calcaneal-Fifth Metatarsal Angle Using Radiograph: A Computer-Aided Diagnosis of Flat Foot for Military New Recruits in Taiwan.

Yang CH, Chou KT, Chung MB, Chuang KS, Huang TC - PLoS ONE (2015)

Bottom Line: The result of this computer-calculated arch angle was compared with manual measurement results from two radiologists, which showed that our automatic arch angle measurement method had a high consistency.In addition, this method had a high accuracy of 97% and 96% as compared with the measurements of radiologists A and B, respectively.The findings indicated that our MI registration measurement method cannot only accurately measure the CA-MT5 angle, but also saves time and reduces human error.

View Article: PubMed Central - PubMed

Affiliation: Department of Radiology, Tao Yuan General Hospital, Ministry of Health and Welfare, Taoyuan City, Taiwan; Department of Biomedical Engineering and Environment Science, National Tsing Hua University, Hsinchu City, Taiwan.

ABSTRACT
Flatfoot (pes planus) is one of the most important physical examination items for military new recruits in Taiwan. Currently, the diagnosis of flatfoot is mainly based on radiographic examination of the calcaneal-fifth metatarsal (CA-MT5) angle, also known as the arch angle. However, manual measurement of the arch angle is time-consuming and often inconsistent between different examiners. In this study, seventy male military new recruits were studied. Lateral radiographic images of their right and left feet were obtained, and mutual information (MI) registration was used to automatically calculate the arch angle. Images of two critical bones, the calcaneus and the fifth metatarsal bone, were isolated from the lateral radiographs to form reference images, and were then compared with template images to calculate the arch angle. The result of this computer-calculated arch angle was compared with manual measurement results from two radiologists, which showed that our automatic arch angle measurement method had a high consistency. In addition, this method had a high accuracy of 97% and 96% as compared with the measurements of radiologists A and B, respectively. The findings indicated that our MI registration measurement method cannot only accurately measure the CA-MT5 angle, but also saves time and reduces human error. This method can increase the consistency of arch angle measurement and has potential clinical application for the diagnosis of flatfoot.

No MeSH data available.


Related in: MedlinePlus

Receiver operating characteristic (ROC) curves of different measurements.(A) ROC curves of radiologist B and MI registration when the measurement of radiologist A was used as the standard. (B) ROC curves of radiologist A and MI registration when the measurement of radiologist B was used as the standard.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4488398&req=5

pone.0131387.g006: Receiver operating characteristic (ROC) curves of different measurements.(A) ROC curves of radiologist B and MI registration when the measurement of radiologist A was used as the standard. (B) ROC curves of radiologist A and MI registration when the measurement of radiologist B was used as the standard.

Mentions: Fig 6 shows the receiver operating characteristic (ROC) curve used to analyze the differences between the measurement obtained by MI registration and those of radiologist A or B when the results measured by radiologists A or B were used as the reference. In Fig 6A, when the measurement of radiologist A was used as the standard, the ROC curve showed that the area under the curve of the ROC (AUC ROC) of radiologist B was 81.4%, with a sensitivity of 64.7% and a specificity of 98.1%; and the AUC ROC of MI registration was 93.2%, with a sensitivity of 88.2% and a specificity of 98.1%. When the measurement of radiologist B was used as the standard, the AUC ROC of radiologist A was 94.2%, with a sensitivity of 94.1% and a specificity of 94.3%; and the AUC ROC of MI registration was 97.6%, with a sensitivity of 100% and a specificity of 95.1%.


Automatic Detection of Calcaneal-Fifth Metatarsal Angle Using Radiograph: A Computer-Aided Diagnosis of Flat Foot for Military New Recruits in Taiwan.

Yang CH, Chou KT, Chung MB, Chuang KS, Huang TC - PLoS ONE (2015)

Receiver operating characteristic (ROC) curves of different measurements.(A) ROC curves of radiologist B and MI registration when the measurement of radiologist A was used as the standard. (B) ROC curves of radiologist A and MI registration when the measurement of radiologist B was used as the standard.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4488398&req=5

pone.0131387.g006: Receiver operating characteristic (ROC) curves of different measurements.(A) ROC curves of radiologist B and MI registration when the measurement of radiologist A was used as the standard. (B) ROC curves of radiologist A and MI registration when the measurement of radiologist B was used as the standard.
Mentions: Fig 6 shows the receiver operating characteristic (ROC) curve used to analyze the differences between the measurement obtained by MI registration and those of radiologist A or B when the results measured by radiologists A or B were used as the reference. In Fig 6A, when the measurement of radiologist A was used as the standard, the ROC curve showed that the area under the curve of the ROC (AUC ROC) of radiologist B was 81.4%, with a sensitivity of 64.7% and a specificity of 98.1%; and the AUC ROC of MI registration was 93.2%, with a sensitivity of 88.2% and a specificity of 98.1%. When the measurement of radiologist B was used as the standard, the AUC ROC of radiologist A was 94.2%, with a sensitivity of 94.1% and a specificity of 94.3%; and the AUC ROC of MI registration was 97.6%, with a sensitivity of 100% and a specificity of 95.1%.

Bottom Line: The result of this computer-calculated arch angle was compared with manual measurement results from two radiologists, which showed that our automatic arch angle measurement method had a high consistency.In addition, this method had a high accuracy of 97% and 96% as compared with the measurements of radiologists A and B, respectively.The findings indicated that our MI registration measurement method cannot only accurately measure the CA-MT5 angle, but also saves time and reduces human error.

View Article: PubMed Central - PubMed

Affiliation: Department of Radiology, Tao Yuan General Hospital, Ministry of Health and Welfare, Taoyuan City, Taiwan; Department of Biomedical Engineering and Environment Science, National Tsing Hua University, Hsinchu City, Taiwan.

ABSTRACT
Flatfoot (pes planus) is one of the most important physical examination items for military new recruits in Taiwan. Currently, the diagnosis of flatfoot is mainly based on radiographic examination of the calcaneal-fifth metatarsal (CA-MT5) angle, also known as the arch angle. However, manual measurement of the arch angle is time-consuming and often inconsistent between different examiners. In this study, seventy male military new recruits were studied. Lateral radiographic images of their right and left feet were obtained, and mutual information (MI) registration was used to automatically calculate the arch angle. Images of two critical bones, the calcaneus and the fifth metatarsal bone, were isolated from the lateral radiographs to form reference images, and were then compared with template images to calculate the arch angle. The result of this computer-calculated arch angle was compared with manual measurement results from two radiologists, which showed that our automatic arch angle measurement method had a high consistency. In addition, this method had a high accuracy of 97% and 96% as compared with the measurements of radiologists A and B, respectively. The findings indicated that our MI registration measurement method cannot only accurately measure the CA-MT5 angle, but also saves time and reduces human error. This method can increase the consistency of arch angle measurement and has potential clinical application for the diagnosis of flatfoot.

No MeSH data available.


Related in: MedlinePlus