Unmet need
Pancreatic ductal adenocarcinoma (PDAC) is the most frequent form of pancreatic cancer, with a survival rate of less than 10% at five years largely due to late-stage diagnosis1. Non-small cell lung cancer (NSCLC) is the most common type of lung cancer, which is the leading cause of cancer death worldwide, primarily due to metastatic disease2. Although routine clinical imaging methods (CT/MRI) represent the current standard of care for at-risk individuals in pancreatic and lung cancer surveillance, the cost of imaging, the availability of imaging facilities, the risks associated with radiation or contrast agents, all represent potential obstacles in expanding the imaging-based approaches to meet the demands of population level screening for PDAC or lung cancer. There is an unmet need for a simple, convenient and cost-effective screening tool that can offer high sensitivity and specificity for detecting cancer at population levels.
Technology overview
Johns Hopkins researchers have developed a novel data-driven approach using a system of artificial neural networks (ANNs) to process high resolution 1H MR spectra of human plasma or serum samples to distinctively identify individuals with undetected cancer from those who are healthy or may have other confounding forms of pancreatic or lung diseases to avoid false detection which is also of paramount importance in cancer screening. Since the spectral features obtained from plasma or serum reflect intricate metabolic processes associated with cancer or other disease processes, the ANN system can be extensively and optimally trained to track and discriminate subtle changes in the features with high sensitivity and specificity. This is the first of its kind approach in that it solely relies on the spectral data to learn the distinguishing patterns in its entirety from the training samples, so that the approach can be well adapted to diverse population demographics for detecting PDAC or NSCLC. As an AI-enabled biofluid-based screening tool, the approach can be extended, with availability of samples, to other cancers where early detection may play a crucial role in long term survival and also to make use of other types of biofluids other than plasma or serum from which 1H MR spectra can be obtained.
Stage of development
Proof of concept. The implementation and rigorous testing were performed for PDAC, and in a limited way for NSCLC pending acquisition of more samples for training.
References:
1. Vincent, A., Herman, J., Schulick, R., Hruban, R.H. & Goggins, M. Pancreatic cancer. Lancet 378, 607-620 (2011).
2. Herbst, R.S., Morgensztern, D. & Boshoff, C. The biology and management of non-small cell lung cancer. Nature 553, 446-454 (2018).
