Register for free to
view the full innovation profile.
AI to identify whether lung nodules are benign or malignant, and using radiogenomics to infer the tumour genetics of malignant neoplasms.
About
Our innovation is unique in the field of AI-based image recognition in healthcare. We give the size, location in three dimensions, and a percent likelihood of benign versus malignant for the lung nodules identified from CT chest scans. We do, though, have further unique advantages. Tumours can exhibit marked genetic temporal and spatial heterogeneity. The current biopsy techniques are only able to represent a tiny portion of this, providing limited information for the treating oncologist. By exploiting radiogenomics, we will be able to give the mutation status of the entire tumour from the CT scan. This will provide more personalised treatment options for patients, allow a more rationale use of chemotherapies, and will provide further research insights into therapeutic target developments. To further add to our product, we are incorporating natural language processing. Clinical medicine works because doctors are able to explain their diagnoses to both their colleagues and to their patients, and we feel systems that exploit artificial intelligence should be no different. Currently in development is a means to facilitate this need, to allow our system to explain why it gave its diagnosis, and to allow quicker clinical acceptance of this uniquely important diagnostic aid.
Key Benefits
1. We will quickly and accurately be able to flag suspicious lung nodules for further review by the radiologist. Aside from the obvious patient advantages, this will save time for the radiologist, help triage their work flow, and allow them to focus on more cognitively taxing elements of their work. Efficiency and thus revenues will be increased, risks will be mitigated, and costs will be reduced from missed diagnoses. 2. By incorporating radiogenomics, we will provide a more holistic view of the tumour than is currently available. This will provide superior results to current biopsy techniques as well as mitigating the patient risks that can occur during the extraction of the tissue for biopsy. In addition, the oncologists will be able to provide more personalised treatment options to their patients, and the genetic tumour evolution can be monitored throughout treatment in non-resected specimens. 3. Natural language processing and the ability of the system to provide an explanation for its diagnosis is key for increasing clinical acceptance.
Applications
Our innovation is applicable to aiding the diagnosis as well as management of lung cancer. In addition, research insights into therapeutic targets can be gleaned. Future applications to other forms of malignancy can be envisioned.