Logo Logo
Help
Contact
Switch language to German
VOC pattern recognition of lung cancer: a comparative evaluation of canine and eNose- based strategies using different sampling materials
VOC pattern recognition of lung cancer: a comparative evaluation of canine and eNose- based strategies using different sampling materials
Parts of the abstract are already published in Biehl et al., Acta oncologica 2019. Background: Lung cancer is the leading cause of cancer related death. Even with optimal treatment, 80-90% of lung cancer patients die within 5 years. However, the prognosis can be significantly improved if the disease is detected and treated in early stages. In recent years, it has been reported that dogs are able to detect (lung) cancer. It has also been reported that an electronic nose (eNose) with chemical sensors can detect volatile organic compounds (VOC) via VOC pattern recognition. Both, dogs and eNose, could help to identify tumors in their early stages by screening. Aims: In this study the following analysis were performed: 1. Comparison of the results of experienced working dogs versus family dogs regarding the detection capability of non-cancer breath samples (specificity) and cancer breath samples (sensitivity) in order to understand how to optimize training. 2. Influence of different breath sample carrier materials on the results achieved by the dogs. 3. Comparison of breath samples by eNose, for those collected and directly assessed in respiratory bags, with those collected on carrier materials and assessed at a later time. 4. Comparison of results achieved by dogs with results by eNose. 5. Development of a strategy for a volatile profiling by dogs and eNose using a suitable carrier material. Material and methods: In the first part of the study, using a methodological approach, two dog teams were employed. Dog Team 1 worked with 5 experienced working dogs, while dog team 2 trained 5 ordinary family dogs with no prior work experience, to discover which dogs were better qualified and the best training method. To find the best carrier material for breath sampling, we compared charcoal filled glass tubes with fleece based earloop masks stored in plastic cups. Breath samples were collected at the Asklepios Klinik Gauting; 70 cancer breath samples from patients with malignant lung disease, and 88 control breath samples from healthy subjects. In the second part of the study, 5 experienced working dogs were trained with revised and improved training methods learning from experiences in the first part. Two fleece-based carrier materials were selected for breath sample collection: a) glass tubes containing two different (siliconized and non-siliconized) fleeces and the previously tested fleece earloop masks. Testing was done by the dog group on the fleeces in glass tubes and by eNose on both breath sample carrier materials. 9 breath samples from patients with lung cancer, as well as 35 control breath samples from COPD patients and healthy volunteers were taken. Results: In the first part of the study it was shown overall that experienced working dogs performed better than family dogs and the dogs achieved a sensitivity of 45-59% and a specificity of 45-69%. Charcoal based breath sample carrier materials did not qualify for detection of VOC by dogs. In the second part of the study, the dogs achieved a specificity of 83% and a sensitivity of 56%, but with considerable differences between individual dogs. The eNose provided a specificity of 97% for both fleece based carrier materials and a sensitivity of 89% for fleece filled glass tubes and 100% for earloop masks. Measurements of breath samples collected directly in respiratory bags as reference measurements achieved a sensitivity and specificity of 100%. Conclusion: Our data shows that both experienced working dogs as well as family dogs have the potential to distinguish between breath samples from cancer patients and non-cancer samples. Experienced working dogs can be trained more easily, faster, and achieved better results. However, the accuracy of the dogs depended very much on; the type of training; the performance of the individual dog; and the carrier material used. A comparison of the dogs’ results to those of eNose showed better results for both specificity and sensitivity by eNose. Both tested carrier materials, fleeces in glass tubes and fleece based earloop masks, can be successfully used as carrier materials for breath samples. There are many possibilities for further eNose studies such as collecting breath samples with qualified carrier material, storing them for a certain period of time, and sending them to a location where they can then be assessed by eNose.
canine scent detection, electronic Nose, volatile factors, lung cancer
Biehl, Wiebke
2021
English
Universitätsbibliothek der Ludwig-Maximilians-Universität München
Biehl, Wiebke (2021): VOC pattern recognition of lung cancer: a comparative evaluation of canine and eNose- based strategies using different sampling materials. Dissertation, LMU München: Faculty of Medicine
[thumbnail of Biehl_Wiebke.pdf]
Preview
PDF
Biehl_Wiebke.pdf

1MB

Abstract

Parts of the abstract are already published in Biehl et al., Acta oncologica 2019. Background: Lung cancer is the leading cause of cancer related death. Even with optimal treatment, 80-90% of lung cancer patients die within 5 years. However, the prognosis can be significantly improved if the disease is detected and treated in early stages. In recent years, it has been reported that dogs are able to detect (lung) cancer. It has also been reported that an electronic nose (eNose) with chemical sensors can detect volatile organic compounds (VOC) via VOC pattern recognition. Both, dogs and eNose, could help to identify tumors in their early stages by screening. Aims: In this study the following analysis were performed: 1. Comparison of the results of experienced working dogs versus family dogs regarding the detection capability of non-cancer breath samples (specificity) and cancer breath samples (sensitivity) in order to understand how to optimize training. 2. Influence of different breath sample carrier materials on the results achieved by the dogs. 3. Comparison of breath samples by eNose, for those collected and directly assessed in respiratory bags, with those collected on carrier materials and assessed at a later time. 4. Comparison of results achieved by dogs with results by eNose. 5. Development of a strategy for a volatile profiling by dogs and eNose using a suitable carrier material. Material and methods: In the first part of the study, using a methodological approach, two dog teams were employed. Dog Team 1 worked with 5 experienced working dogs, while dog team 2 trained 5 ordinary family dogs with no prior work experience, to discover which dogs were better qualified and the best training method. To find the best carrier material for breath sampling, we compared charcoal filled glass tubes with fleece based earloop masks stored in plastic cups. Breath samples were collected at the Asklepios Klinik Gauting; 70 cancer breath samples from patients with malignant lung disease, and 88 control breath samples from healthy subjects. In the second part of the study, 5 experienced working dogs were trained with revised and improved training methods learning from experiences in the first part. Two fleece-based carrier materials were selected for breath sample collection: a) glass tubes containing two different (siliconized and non-siliconized) fleeces and the previously tested fleece earloop masks. Testing was done by the dog group on the fleeces in glass tubes and by eNose on both breath sample carrier materials. 9 breath samples from patients with lung cancer, as well as 35 control breath samples from COPD patients and healthy volunteers were taken. Results: In the first part of the study it was shown overall that experienced working dogs performed better than family dogs and the dogs achieved a sensitivity of 45-59% and a specificity of 45-69%. Charcoal based breath sample carrier materials did not qualify for detection of VOC by dogs. In the second part of the study, the dogs achieved a specificity of 83% and a sensitivity of 56%, but with considerable differences between individual dogs. The eNose provided a specificity of 97% for both fleece based carrier materials and a sensitivity of 89% for fleece filled glass tubes and 100% for earloop masks. Measurements of breath samples collected directly in respiratory bags as reference measurements achieved a sensitivity and specificity of 100%. Conclusion: Our data shows that both experienced working dogs as well as family dogs have the potential to distinguish between breath samples from cancer patients and non-cancer samples. Experienced working dogs can be trained more easily, faster, and achieved better results. However, the accuracy of the dogs depended very much on; the type of training; the performance of the individual dog; and the carrier material used. A comparison of the dogs’ results to those of eNose showed better results for both specificity and sensitivity by eNose. Both tested carrier materials, fleeces in glass tubes and fleece based earloop masks, can be successfully used as carrier materials for breath samples. There are many possibilities for further eNose studies such as collecting breath samples with qualified carrier material, storing them for a certain period of time, and sending them to a location where they can then be assessed by eNose.