Summary: Kience presented the successful results of its first validation study for a less-invasive test that uses blood and urine samples to screen for up to 42 solid tumors and five hematological malignancies, leveraging a machine learning algorithm and the Barcelona Criteria to enhance accuracy.
Takeaways:
- Kience’s non-invasive test can detect up to 42 solid tumors and five blood cancers, offering early screening before symptoms appear, with a focus on people aged 40 and over.
- The test uses a proprietary machine learning algorithm, incorporating the Barcelona Criteria to improve accuracy by reducing false positives and overdiagnosis errors.
- A larger international clinical trial with 26,000 patients is scheduled to run from September 2024 to December 2026 to further refine the test’s accuracy.
Kience presented the results of the first validation study of its non-invasive test, which uses blood and urine samples to provide early and accurate screening for up to 42 different types of solid tumors and five hematological malignancies (even in the early stages), before the first symptoms and signs appear.
Testing with Machine Learning
The test, especially aimed at people aged 40 and over, is claimed to be much less invasive, more economical, and more guaranteed to offer fewer false positives than other tests currently on the market to detect different types of cancer, due to it being based on a machine learning algorithm of its own creation that includes, among other innovations, the Barcelona Criteria, four criteria established by Rafael Molina, MD, PhD, together with other members of the Spanish Society of Clinical Chemistry (SEQCML) in 1994, on the correct interpretation of tumor markers.
Developing the Algorithm
For the development of the algorithm, Kience initially used a sample size (n) of 192,090 patients, which is the sample resulting from the agglutination of more than 1,600 previous studies (some of them acquired through different technology transfer contracts with the developing medical centers). It was then processed using complex mathematical functions, mainly serial and parallel approximations, to increase specificity and sensitivity, respectively. From here, a sample of 314 patients (166 in the final sample) was recruited to validate the efficacy of the test. The blood and urine samples of these patients were analyzed by Laboratorio Echevarne (Sant Cugat del Vallès, Barcelona), and subsequently processed with the algorithm for the evaluation of their results.
Without the application of the Barcelona Criteria, the algorithm achieved a sensitivity of 100% but a specificity of only 67%, with a positive predictive value (PPV) of 21%. In contrast, when applying the Barcelona Criteria, the specificity was 100% and the PPV improved considerably, also reaching 100%. Even so, it is important to note that these results were obtained with a final sample of 166 patients (148 patients were excluded from the trial because they did not meet any of the trial criteria).
Further reading: New Blood Test Enables Minimally Invasive Glioma Diagnosis
Nevertheless, the results themselves demonstrate how important it is to include the Barcelona Criteria in any test that makes use of tumor markers (such as cancer screening tests), in order to avoid false positives and, consequently, overdiagnosis errors.
The company has launched a new international clinical trial with 26,000 patients to further improve the accuracy of the algorithm in the general population. This new trial will run from September 2024 to December 2026.
Featured image: Kience booth at the ESMO 2024 Congress with the poster of the results presented. Photo: Kience
