The cancerous diseases that are a priority for our research are ovarian cancer and lung cancer.
Ovarian cancer is one of the most difficult diseases to detect in the general population. According to statistics, it is detected in the early stage in just some 25 % of patients. Approximately 90 % of those patients can then benefit from the early diagnosis in the form of conventional surgical removal and chemotherapy. However, most patients come to their doctor with a more advanced form of tumour.
In real laboratory practice, when detecting an ovarian tumour it is hard to achieve sufficient diagnostic quality to achieve a figure of 100 % for the four parameters (sensitivity, specificity, NPV and PPV). Laboratory diagnostic tests used to detect ovarian carcinoma are based, amongst other things, on the concentration of a substance called CA-125. This is a molecule that is detected in the patient’s serum and if it exceeds a certain level it indicates that a tumour could be present in the body. If CA-125 is monitored separately without any other similar substances, its values are approximately as follows: sensitivity – 79%; specificity – 76%, PPV – 63%; NPV – 87% (http://doi.org). Its low specificity prevents this molecule from being used to detect the early stage of the disease.
Lung cancer is the third most common form of cancer within the Czech Republic. Given the fact that there are no early warning signals that would enable it to be detected in its earliest stages, this disease tends to be diagnosed at a late stage. Our research therefore also includes training dogs to detect the early stages of this disease.
Owing to the above we have decided to conduct a series of experiments, using dogs that are trained to tag samples containing the odour of “ovarian tumours” or tumours of the lungs. We have defined some hypotheses that we would like to gradually verify. We have also started collaborating with specialists in various fields, including oncology, gynaecology, pulmonary medicine, epidemiology and quality systems. Naturally, we also collaborate with cynology specialists.
The first results of training dogs in the home environment are shown in the table below. These are the ongoing results of the subjective assessment of cynologists without blind trials. We are aware that given that the tests are unblind, the results may be overstated. A double-blind study is planned for the near future. We will be happy to keep you informed.
If you are interested in our project, you can join in, either as a volunteer or as a specialist, who would like to work with us primarily in obtaining samples from cancer patients. Please contact us. We will be very happy to have you involved.
Pilot results of diagnostic dog tests (unblind tests)
Sensitivity expresses the percentage of correct positive results from the given set.
Sensitivity expresses the percentage of correct negative results from the given set.
Positive predictive value – PPV expresses the test’s ability to find a sick person amongst all individuals with a positive test result. This is the likelihood that, if the test is positive, the given person really is ill.
Or otherwise PPV expresses the percentage likelihood that the patient will have the disease that is being sought, if the test result is positive.
Negative predictive value – NPV expresses the test’s ability to find healthy people amongst all individuals with a negative test result. This is therefore the likelihood that a person with a negative test really is healthy.
In other words, NPV expresses the percentage likelihood that the patient will not have the disease that is being sought, if the test result is negative.
Source: Klinická biochemie (Clinical Biochemistry) (second, reworked edition). Jaroslav Racek et al., 2006, ISBN80-7262-324-9
What we have planned
We will be extending our training. The next disease we will soon be focusing our research on is prostate cancer. This is disease is one of the most common causes of death amongst men. Unfortunately, laboratory examinations of the molecule known under the abbreviation PSA cannot be used for large-scale screening due to their low specificity. This can result in unnecessary cases of prostate biopsies. The aim of the training will be to teach dogs to distinguish between urine samples from patients with confirmed benign prostatic hyperplasia from the urine samples of patients with the early stages of prostate cancer.